Crystal structure

Abstract

This invention relates to a crystallisable composition comprising a TSHR polypeptide, to crystals comparing a TSHR polypeptide and to TSHR-related applications.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS

This application is a divisional of U.S. patent application Ser. No. 13/306,496, filed Nov. 29, 2011, which is a divisional of U.S. patent application Ser. No. 11/896,073, filed Aug. 29, 2007, now U.S. Pat. No. 8,097,699, which claims priority to U.S. Provisional Patent Application No. 60/840,967, filed on Aug. 30, 2006 and U.S. Provisional Patent Application No. 60/901,685, filed Feb. 16, 2007, the disclosures of each of which are incorporated by reference in their entireties. This application also claims priority to U.K. Patent Application No. GB 0617239.3, filed Aug. 31, 2006 and U.K. Patent Application No. GB 0703070.3, filed Feb. 16, 2007, the disclosure of each of which are herein incorporated by reference in their entireties.

FIELD OF THE INVENTION

The present invention is concerned with the thyrotropin receptor, also known as the Thyroid Stimulating Hormone Receptor, (TSHR) and autoantibodies reactive with the TSHR, and in particular the interactions between the TSHR and such autoantibodies as determined by X-ray crystallography.

BACKGROUND

Thyrotropin or thyroid stimulating hormone (TSH) is a pituitary hormone that regulates thyroid function via the TSHR (Szkudlinski M W, Fremont V, Ronin C, Weintraub 2002 Thyroid-stimulating hormone and TSHR structure-function relationships. Physiological Reviews 82: 473-502). Binding of TSH to the TSHR triggers receptor signaling which leads to stimulation of formation and release of thyroid hormones; thyroxine (T4) and tri-iodothyronine (T3). A feedback mechanism involving the levels of T4 and T3 in the circulation and thyrotropin releasing hormone (TRH) secreted by the hypothalamus controls the release of TSH that in turn controls thyroid stimulation and the levels of thyroid hormones in serum.

The TSHR is a G-protein coupled receptor and has three domains: a leucine rich domain (LRD), a cleavage domain (CD) and a transmembrane domain (TMD) (Núñez Miguel R, Sanders J, Jeffreys J, Depraetere H, Evans M, Richards T, Blundell T L, Rees Smith B, Furmaniak J 2004 Analysis of the thyrotropin receptor-thyrotropin interaction by comparative modeling. Thyroid 14: 991-1011). The TSHR shows amino acid and structural similarities with the other glycoprotein hormone receptors ie luteinizing hormone receptor (LHR) and follicle-stimulating hormone receptor (FSHR). The structure of the FSHR in complex with its ligand (ie FSH) has been solved at 2.9 Å resolution (Fan Q R, Hendrickson W A 2005 Structure of human follicle-stimulating hormone in complex with its receptor. Nature 433: 269-277).

It is well documented in the art that some patients with autoimmune thyroid disease (AITD), which is the most common autoimmune disease affecting different populations worldwide, have autoantibodies reactive with the TSHR (Rees Smith B, McLachlan S M, Furmaniak J 1988 Autoantibodies to the thyrotropin receptor. Endocrine Reviews 9: 106-121). In the majority of cases, these autoantibodies bind to the TSHR and mimic the actions of TSH thereby stimulating the thyroid to produce high levels of T4 and T3. These autoantibodies are described as thyroid stimulating autoantibodies, or TSHR autoantibodies (TRAbs) with stimulating activity or TSH agonist activity. The feedback mechanism which usually controls thyroid function is no longer effective in the presence of thyroid stimulating autoantibodies and the patients present with symptoms of a hyperactive thyroid (excess of thyroid hormones in serum). This condition is known as thyrotoxicosis or Graves' disease. In some patients the TRAbs with stimulating activity are thought to interact with TSHRs in retro-orbital tissues and contribute to causing the eye signs of Graves' disease.

In some patients with AITD, autoantibodies bind to the TSHR, preventing TSH from binding to the receptor but have no ability to stimulate TSHR activity; these types of autoantibodies are known as TRAbs with blocking activity or TSH antagonist, activity.

TSHR autoantibodies when present in serum of pregnant women in high concentrations can cross the placenta and may cause neonatal thyrotroxicosis (in the case of stimulating autoantibodies) or neonatal hypothyroidism (in the case of blocking autoantibodies) (Rees Smith B, McLachlan S M, Furmaniak J 1988 supra).

A human monoclonal autoantibody which acts as a powerful thyroid stimulator (hMAb TSHR1) has been described in detail in International Patent Application WO2004/050708A2. The binding site for hMAb TSHR1 has been found to be located on the surface of the TSHR leucine rich domain (LRD) and overlaps extensively with the binding site for TSH. However, the binding pocket for TSH or hMAb TSHR1 is conformational and involves discontinuous regions of the TSHR folding together. Characterization of the binding site for hMAb TSHR1 in detail, in particular the important contact amino acids in the interaction between the TSHR and hMAb TSHR1, is of critical importance in studies which aim to improve the diagnosis and management of diseases associated with an autoimmune response to the TSHR.

International patent application WO 2006/016121A discloses a mutated TSHR preparation including at least one point mutation which can be used in the differential screening and identification of patient serum stimulating TSHR autoantibodies, patient serum blocking TSHR autoantibodies and thyroid stimulating hormone in a sample of body fluid from a patient which is being screened. The invention described in international patent application number WO2006/016121A provides useful information regarding the regions of the TSHR which are important in the interaction with various antibodies including hMAb TSHR 1, a mouse monoclonal antibody (9D33) with TSHR blocking activity and with TSH. However, details of the interactions between amino acids in the TSHR and amino acids in hMAb TSHR1 at the atomic level could not be derived from even the best experimental studies, such as those described in WO2006/016121A, which involved mutating the TSHR.

The present invention is based on the preparation of a complex formed by a fragment of the TSHR LRD (which is involved in forming the binding pocket for TSH and hMAb TSHR1) and the Fab fragment of hMAb TSHR1. The hMAb TSHR1 preparations described in this specification are referred to as M22 for convenience. M22 IgG can be purchased from RSR Ltd. The TSHR fragment covering amino acids 1-260 (TSHR260) in complex with hMAb TSHR1 (M22) Fab fragment is referred to as TSHR260-M22. A TSHR260-M22 complex was purified, concentrated and crystallized. The data from X-ray diffraction were used to solve the structure of TSHR260 as described in the present invention. The structure of M22 Fab solved at 1.65 Å resolution has been described before (Sanders J, Jeffreys J, Depraetere H, Evans M, Richards T, Kiddie A, Brereton K, Premawardhana L D K E, Chirgadze D Y, Núñez Miguel R, Blundell T L, Furmaniak J, Rees Smith B 2004 Characteristics of a human monoclonal autoantibody to the thyrotropin receptor: sequence structure and function. Thyroid 14: 560-570). The structure of M22 in the complex was compared to that of un-bound M22. The interactions between TSHR260 and M22 were then refined at the atomic level.

To date highly purified TSHR preparations with their TSH and TRAb binding activity intact have not been available. The purified TSHR preparations described in the art were denatured in part and not pure or homogenous enough for crystallisation.

SUMMARY OF THE INVENTION

According to one aspect of the invention there is provided a crystallisable composition comprising a TSHR polypeptide, that is to say a polypeptide comprising contiguous amino acids from the primary sequence of a thyroid stimulating hormone receptor.

According to another aspect of the invention there is provided a crystal comprising a TSHR polypeptide.

According to another aspect of the invention there is provided a crystallisable complex comprising a TSHR polypeptide and a TSHR-binding entity.

Such a complex is advantageous in that the TSHR-binding entity may stabilise the TSHR polypeptide. The invention also provides methods of producing crystallisable complexes comprising a TSHR polypeptide and a TSHR-binding entity in which the TSHR-binding entity has a relatively high affinity for the TSHR polypeptide, compared to TSH, and stabilises the TSHR polypeptide. A suitable TSHR-binding entity is M22. A TSHR binder may be used to stabilise the TSHR polypeptide as it is produced. For example, in production of a TSHR polypeptide in cells, such as insect cells, expressing a TSHR-polypeptide encoding DNA construct, a TSHR binder, such as M22 Fab can be added to the cells to “capture” and stabilise the TSHR polypeptide as it is secreted.

According to another aspect of the invention there is provided a method of producing a complex of a TSHR polypeptide and a TSHR binder by expression of a DNA construct encoding the TSHR polypeptide, the method comprising expressing the TSHR polypeptide and adding the TSHR binder to stabilise the secreted TSHR polypeptide.

According to a further aspect of the invention there is provided a method of producing a complex of a TSHR polypeptide and a TSHR binder by expression of a DNA construct encoding the TSHR polypeptide linked to a TSHR binding entity such as M22 Fab, the method comprising expressing the TSHR polypeptide linked to a TSHR binding entity.

According to another aspect of the invention there is provided a co-crystal comprising a TSHR polypeptide and the TSHR binding entity.

The TSHR polypeptide preferably comprises a mammalian TSHR sequence. Preferably, the mammalian sequence is of human origin, but the use of chimpanzee, African green monkey, rhesus monkey, canine, feline, porcine, equine, bovine or guinea pig sequences is also contemplated. Preferably the TSHR polypeptide includes at least a portion of the leucine-rich domain of TSHR, most preferably a human sequence. Preferably, the TSHR polypeptide includes amino acids 22-260 of the wild-type human TSHR sequence. The TSHR polypeptide in a complex according to the invention may comprise the full wild-type sequence of TSHR. The TSHR polypeptide, alternatively, may include mutations of the wild-type sequence. For example specific amino acids in the wild-type sequence may be replaced with alternative amino acids. These substitutions may be conservative, that is to say replacing one amino acid residue with another amino acid having similar properties.

The TSHR-binding entity may be an antibody or a portion thereof. A suitable antibody may be an autoantibody or a portion thereof or a TSHR binding entity derived therefrom. Suitable antibodies include monoclonal antibodies. A suitable antibody portion is M22 Fab. According to another aspect of the invention there is provided a co-crystal comprising a crystalline form of the TSHR polypeptide having coordinates, as determined by X-ray crystallography, of FIG. 9a or 9b.

The analysis of co-crystals in accordance with the invention has provided atomic coordinates and structure factors of M22 in complex with TSHR260 at 2.55 Å resolution. Such a level of confidence can be only obtained from X-ray crystallographic analysis of the structure of the complex formed by the two molecules (the TSHR and M22) solved at a high resolution. This is advantageous because only a crystal structure analysis of the complex between the two molecules, as compared with other methods of predicting amino acids which are important for binding, allows the identification of interactions (including the distance between the atoms of the residues involved). Furthermore, the complex interactions between the receptor and ligand can only be obtained from the crystal structure.

The information provided by the crystal structure of M22 in complex with the TSHR is surprising. In particular, the information was not available before, and studies such as modeling and mutation experiments only provided rudimentary hints as to amino acids which might be involved in interactions between the TSHR and TSH and the TSHR and TSHR autoantibodies. All these earlier studies could show was that there was extensive overlap between the TSHR binding sites for TSH and TSHR autoantibodies. It was also clear that the TSHR and FSHR were closely related structurally. However, there was no indication of the whole complexity of the interactions between a TSHR autoantibody such as M22 and the TSHR or of the actual true detailed structure of the TSHR LRD.

According to another aspect of the invention there is provided a machine-readable data storage medium encoded with data relating to at least a portion of the coordinates of TSHR polypeptide amino acids of FIG. 9a or 9b or a homologue thereof. Preferably, the data includes all of the TSHR polypeptide amino acid coordinates of FIG. 9a or 9b.

According to another aspect of the invention there is provided a computer system for presenting a representation of a three-dimensional structure of a TSHR polypeptide, or a homologue of such a TSHR polypeptide, in which the homologue has a root mean square deviation from the backbone atoms of between 0 Å and 4 Å, the computer system including data storage means including data corresponding to TSHR polypeptide amino acid coordinates of FIG. 9a or 9b. The computer system may include data storage means including data corresponding to coordinates of a chemical entity interacting with the TSHR polypeptide or homologue thereof.

Preferably the computer system is arranged to provide a representation of a three-dimensional structure of the TSHR polypeptide or homologue thereof interacting with a chemical entity. The chemical entity may be an antibody or a portion thereof. Preferably, the antibody portion is M22 Fab.

The computer system may include a display for displaying a representation of the three-dimensional structure of the TSHR polypeptide. Preferably, the computer system is arranged to display the chemical entity interacting with the TSHR polypeptide or homologue thereof.

According to another aspect of the invention there is provided an electronic representation of a three-dimensional structure of a TSHR polypeptide. Preferably the TSHR polypeptide includes the leucine rich domain of human TSHR. More preferably, the representation represents at least amino acids 30-257 of human TSHR. According to another aspect of the invention there is provided an electronic representation of the three-dimensional structure of the TSHR polypeptide and an antibody thereto or a portion thereof.

According to another aspect of the invention there is provided a method of identifying a chemical entity which will interact with at least one amino acid of a TSHR polypeptide three-dimensional structure or homologue thereof according to a representation provided by computer system, or as represented by an electronic representation, according to a previous aspect of the invention. The chemical entity may be a TSHR agonist or antagonist. A chemical entity may be identified which will interact by forming a hydrogen bond with the least one of the TSHR amino acids: K129, E107, K58 and Y185. Additionally or alternatively a chemical entity may be identified which will interact by forming van der Waals interactions with at least one of the TSHR amino acid residues R255, R80, K129, R38 and K183. Additionally or alternatively a chemical entity may be identified which will interact by forming electrostatic interactions with at least one of the TSHR amino acid residues D151, K58, K129, R80, K209, K183. Additionally or alternatively a chemical entity is identified which will interact by forming ion pair interactions with TSHR amino acid residue K209.

According to a further aspect of the invention there is provided a method of identifying a chemical entity which may potentially interfere with the binding of autoantibodies to the TSHR, the method comprising identifying a chemical which interacts with the highly positively charged ridge at the N-terminal end of the concave surface of the TSHR leucine-rich domain. The autoantibodies may be thyroid stimulating autoantibodies, or TSH antagonists i.e. TSH autoantibodies with blocking activity. A suitable chemical entity may interact with at least one of the following TSHR amino acids: R38, K58, R80, H105 and K129.

According to the further aspect of the invention there is provided a method of identifying a chemical entity which may potentially interfere with the binding of autoantibodies to the TSHR, the method comprising identifying a chemical entity which at least substantially fills a negatively charged cavity on the M22 surface formed by M22 hypervariable regions H1, H2 and H3 (FIG. 5) using a computer or an electronic representation according to a previous aspect of the invention. The autoantibodies may be thyroid stimulating autoantibodies, or TSH antagonists i.e. TSH autoantibodies with blocking activity.

Typically such a method comprises identifying a chemical entity which will at least substantially disrupt a thyroid stimulating autoantibody binding to the TSHR residue R255. Additionally or alternatively the method comprises identifying a chemical entity which will disrupt a thyroid stimulating autoantibody binding to amino acid residue K209 of human TSHR.

In methods of the invention, the potential interaction of a chemical entity which has been identified as binding to a TSHR polypeptide, or interfering with the binding to a TSHR polypeptide of a TSHR binder, with other receptors is determined. One form of potential interaction is binding. Other receptors may include follicle stimulating hormone receptor or luteinizing hormone receptor.

According to another aspect of the invention there is provided a method of detecting autoantibodies to TSHR including comparing binding of a putative autoantibody and a chemical entity identified as interacting with a TSHR polypeptide by a method in accordance with the invention with a TSHR polypeptide.

The invention in its various aspects is advantageous in that it allows the skilled addressee to design new pharmaceutical compositions that will specifically prevent thyroid stimulating autoantibodies binding to the TSHR thereby providing new treatments for autoimmune conditions such as Graves' disease. The invention also allows the skilled addressee to design new pharmaceutical compositions that will specifically stimulate tissues containing the TSHR.

According to a further aspect of the invention there is provided a chemical entity identified by a method according to a preceding aspect of the invention. According to a further aspect of the invention there is provided a chemical compound including at least one such chemical entity. According to a further aspect of the invention there is provided a pharmaceutical composition comprising such a chemical compound and a pharmaceutically acceptable carrier.

Such a pharmaceutical composition may be suitable for use in the treatment of Autoimmune Thyroid Disease. Alternatively, such a pharmaceutical composition may be suitable to use in the treatment of Graves' disease. There is also provided a pharmaceutical composition for use in stimulating tissues containing the TSHR.

A chemical entity provided by the invention may be suitable for use in detection of autoantibodies to the TSHR.

According to a further aspect of the invention there is provided the use of a chemical entity or a chemical compound according to a previous aspect of the invention in the preparation of a medicament for the treatment of Autoimmune Thyroid Disease.

According to another aspect of the invention there is provided the use of a chemical entity or a chemical compound according to a previous aspect of the invention in the preparation of a medicament for the treatment of Graves' disease.

Pharmaceutical compositions of this invention comprise any of the compounds of the present invention, and pharmaceutically acceptable salts thereof, with any pharmaceutically acceptable carrier, adjuvant or vehicle. Pharmaceutically acceptable carriers, adjuvants and vehicles that may be used in the pharmaceutical compositions of this invention include, but are not limited to, ion exchangers, alumina, aluminum stearate, lecithin, serum proteins, such as human serum albumin, buffer substances such as phosphates, glycine, sorbic acid, potassium sorbate, partial glyceride mixtures of saturated vegetable fatty acids, water, salts or electrolytes, such as protamine sulfate, disodium hydrogen phosphate, potassium hydrogen phosphate, sodium chloride, zinc salts, colloidal silica, magnesium trisilicate, polyvinyl pyrrolidone, cellulose-based substances, polyethylene glycol, sodium carboxymethylcellulose, polyacrylates, waxes, polyethylene-polyoxypropylene-block polymers, polyethylene glycol and wool fat.

The pharmaceutical compositions of this invention may be administered orally, parenterally, by inhalation spray, topically, by eyedrops or gels, rectally, nasally, buccally, vaginally or via an implanted reservoir. We prefer oral administration or administration by injection. The pharmaceutical compositions of this invention may contain any conventional non-toxic pharmaceutically-acceptable carriers, adjuvants or vehicles. The term “parenteral” as used herein includes subcutaneous, intracutaneous, intravenous, intramuscular, intra-articular, intrasynovial, infrasternal, intrathecal, intralesional and intracranial injection or infusion techniques.

The pharmaceutical compositions may be in the form of a sterile injectable preparation, for example, as a sterile injectable aqueous or oleaginous suspension. This suspension may be formulated according to known techniques using suitable dispersing or wetting agents (such as, for example, Tween 80) and suspending agents. The sterile injectable preparation may also be a sterile injectable solution or suspension in a non-toxic parenterally-acceptable diluent or solvent, for example, as a solution in 1,3-butanediol. Among the acceptable vehicles and solvents that may be employed are mannitol, water, Ringer's solution and isotonic sodium chloride solution. In addition, sterile, fixed oils are conventionally employed as a solvent or suspending medium. For this purpose, any bland fixed oil may be employed including synthetic mono- or diglycerides. Fatty acids, such as oleic acid and its glyceride derivatives are useful in the preparation of injectables, as are natural pharmaceutically-acceptable oils, such as olive oil or castor oil, especially in their polyoxyethylated versions. These oil solutions or suspensions may also contain a long-chain alcohol diluent or dispersant such as Ph. Hely or a similar alcohol.

The pharmaceutical compositions of this invention may be orally administered in any orally acceptable dosage form including, but not limited to, capsules, tablets, and aqueous suspensions and solutions. In the case of tablets for oral use, carriers which are commonly used include lactose and corn starch. Lubricating agents, such as magnesium stearate, are also typically added. For oral administration in a capsule form, useful diluents include lactose and dried corn starch. When aqueous suspensions are administered orally, the active ingredient is combined with emulsifying and suspending agents. If desired, certain sweetening and/or flavouring and/or colouring agents may be added.

The pharmaceutical compositions of this invention may also be administered in the form of suppositories for rectal administration. These compositions can be prepared by mixing a compound of this invention with a suitable non-irritating excipient which is solid at room temperature but liquid at the rectal temperature and therefore will melt in the rectum to release the active components. Such materials include, but are not limited to, cocoa butter, beeswax and polyethylene glycols.

Topical administration of the pharmaceutical compositions of this invention is especially useful when the desired treatment involves areas or organs readily accessible by topical application. For application topically to the skin, the pharmaceutical composition should be formulated with a suitable ointment containing the active components suspended or dissolved in a carrier. Carriers for topical administration of the compounds of this invention include, but are not limited to, mineral oil, liquid petroleum, white petroleum, propylene glycol, polyoxyethylene polyoxypropylene compound, emulsifying wax and water. Alternatively, the pharmaceutical composition can be formulated with a suitable lotion or cream containing the active compound suspended or dissolved in a carrier. Suitable carriers include, but are not limited to, mineral oil, sorbitan monostearate, polysorbate 60, cetyl esters wax, cetearyl alcohol, 2-octyldodecanol, benzyl alcohol and water. The pharmaceutical compositions of this invention may also be applied to the lower intestinal tract in a suitable enema formulation. Topically-transdermal patches are also included in this invention.

The pharmaceutical compositions of this invention may be administered by nasal aerosol or inhalation. Such compositions are prepared according to techniques well-known in the art of pharmaceutical formulation and may be prepared as solutions in saline, employing benzyl alcohol or other suitable preservatives, absorption promoters to enhance bioavailability, fluorocarbons, and/or other solubilizing or dispersing agents known in the art.

The pharmaceutical composition of this invention may be administered as eyedrops, an eye gel or an eye ointment. In the case of eyedrops or eye gels, suitable excipients include, but are not limited to, water, tonicity-modifying agents (e.g., sodium chloride), preservatives (e.g., benzalkonium chloride) and/or buffering agents (e.g., sodium dihydrogen phosphate monohydrate and/or anhydrous disodium phosphate). In the case of eye ointments, suitable excipients include, but are not limited to, white and/or yellow soft paraffin, lanolin and/or liquid paraffin.

The invention allows the skilled addressee to design new ways of measuring and assessing autoantibodies to the TSHR. Current TSHR preparations produced by recombinant DNA technology which are used in some current methods of measuring and assessing autoantibodies to the TSHR are relatively expensive. In addition, the currently used preparations of TSHR are insoluble in water and require the presence of detergents, are relatively “crude” ie they contain a mixture of other proteins and represent a mixture of denatured and non-denatured receptors. A synthetic TSHR polypeptide composition (without the complex seven membrane spanning section of the TSHR ie water soluble and easy to handle) may be designed with the TRAb binding properties based on the interactions found in the complex according to the invention. This composition may be used to develop sensitive, isotopic or non-isotopic assays to measure TRAbs. These new assays can be based on inhibition of binding of M22 (or a different TSHR monoclonal autoantibody or a mixture of TSHR monoclonal autoantibodies or compositions derived from them) or could be based on direct binding to the composition. The composition could be labeled (with isotopic or non-isotopic labels) or conjugated to various reagents known in the art. The composition could be coated onto a solid support (beads, plates, tubes) or used in solution in precipitation assays.

Conversely, a synthetic TSHR binding composition can be designed to be used in inhibition-type assays to replace M22 (or other TSHR monoclonal antibodies) or TSH. M22 IgG or TSH currently used in these assays is relatively expensive to produce and purify. M22 IgG fragments (such as Fab or F(ab′)₂) derived by enzymatic digestion are less stable than IgG, expensive to produce and not easy to handle when labeled with isotopic or non-isotopic labels. The TSHR binding composition that may replace M22 IgG or TSH may be more stable when labeled with isotopic or non-isotopic labels or conjugated to other reagents.

Further, combination of a synthetic composition of the TSHR binding epitope with a synthetic TSHR binder composition could lead to the development of sensitive, easy to produce, easy to use, inexpensive TRAb assays suitable for automation. In addition, the invention allows the skilled addressee to design and test specific amino acid mutations in M22 and in the TSHR that may lead to discovering the amino acids critical for receptor activation.

Furthermore, the invention permits the skilled addressee to understand similarities and differences between the TSHR (involved in thyroid regulation) and the follicle stimulating hormone receptor (FSHR; involved in reproduction). TSHR and FSHR are closely related receptors which have similar structures and bind the ligands TSH and FSH respectively that have a common hormone subunit (a subunit) and show considerable structural similarity themselves. However, TSHR and FSHR have distinct functional activities; the TSHR is involved in thyroid regulation (important for overall body metabolism) while the FSHR is involved in reproduction. The specificity of these receptors for their respective hormones is high and in studies where some low level cross-reactivity has been reported very high molar concentrations of the hormones were used. For the first time the skilled worker can now study the two receptors, compare their structures and the binding arrangements with their respective ligands at the atomic level. For example, comparison of TSHR structure, as disclosed here in, with FSHR structure (Fan & Hendrickson 2005 infra) showed remarkable similarities (only 1.1 Å rmsd on C_α atoms). Also, the binding arrangements between the FSHR and FSH are remarkably similar to that of TSHR and M22. TSH binding arrangements with the TSHR can be now studied (using the available structures and the methods known in the art) and compared to that of FSHR-FSH. TSH or FSH amino acids responsible for their respective specificity can be identified and the evolution of the hormones analyzed using methods known in the art. Further understanding of the 2 closely related receptor-hormone interactions could help the skilled worker to design ligands with the high specificity for the TSHR or the FSHR that would not have undesirable cross-reactivity. Ligands of this type may have applications in the regulation of the reproductive system and for control of thyroid function.

The invention may also allow the skilled addressee to understand the immunological mechanisms which drive the development and production of TSHR autoantibodies in particular and autoantibodies in general.

The invention provides additionally a synthetic water-soluble TSHR polypeptide composition. The polypeptide composition may include amino acids 22-260 of human TSHR.

The data provided by the crystal structure of TSHR260-M22 may allow the design of a binding molecule (such as a polypeptide) that mimics the TSHR binding site for M22 (and also TSHR autoantibodies that have similar surface and binding characteristics). Such a binding molecule can be coupled to an appropriate support and used in the removal of thyroid stimulating autoantibodies. Similar approaches have been used for eliminating autoantibodies to the acetylcholine receptor (Guo C Y et al, Journal of Immunological Methods, 2005, 303, pp 142-147). Accordingly, according to a further aspect of the invention there is provided a method of removing thyroid stimulating hormone receptor antibodies, particularly thyroid stimulating hormone receptor autoantibodies, from a sample containing such thyroid stimulating hormone receptor antibodies, the method comprising providing a binding molecule that includes or mimics a thyroid stimulating hormone receptor binding site for M22, or thyroid stimulating hormone receptor autoantibodies having similar surface and binding characteristics to M22, contacting the sample with the binding molecule whereby thyroid stimulating hormone receptor antibodies bind to the binding molecule and are removed from the sample. A suitable sample may include patient serum containing high levels of thyroid stimulating autoantibodies. To facilitate coupling, the binding molecule may be fused to a neutral protein or other tag that does not affect TSHR autoantibody binding. For example, maltose binding protein can be used as disclosed in Guo C Y (2005) et al, supra. The binding molecule may be coupled to a solid support such as agarose or resin directly or via such a tag. The patient serum containing high levels of thyroid stimulating autoantibodies may then be passed through the immunosorbent (in a process similar to plasma exchange or plasmapheresis) and the serum depleted from the TSHR autoantibodies returned back to the patient. This provides an opportunity for an effective and quick elimination of the autoantibodies responsible for the clinical symptoms of thyroid over-stimulation. This may be of particular importance in the case of a thyroid crisis. Furthermore, elimination of TSHR autoantibodies from the circulation would prevent them from binding to the TSHR in retro-orbital tissue (or other extra-thyroidal sites) thus providing an effective means of controlling severe cases of Graves' ophthalmopathy (or pre-tibial myxoedema). Elimination of TSHR autoantibodies from the circulation of pregnant women would prevent trans-placental passage of the autoantibodies and protect fetal thyroid from the biological effects of autoantibodies. The invention thus also provides a method of treating such conditions, the method including a step of removing thyroid stimulating autoantibodies from a patient having such a condition, or a sample taken from such a patient, by removing the autoantibodies as described immediately above.

BRIEF DESCRIPTION OF THE DRAWINGS

The patent or application file contains at least one drawing executed in color. Copies of this patent or patent application publication with color drawing(s) will be provided by the U.S. Patent and Trademark Office upon request and payment of the necessary fee.

Products and methods in accordance with the invention will now be described by way of example only with reference to the accompanying FIGS. 1-10, in which:

FIG. 1a shows Western blotting analysis of the TSHR260 expressed in High Five cells in the absence or in the presence of M22 IgG. FIG. 1b shows Western blotting analysis of the TSHR277, C-del TSHR and TSHR260 in the culture supernatants from High Five cells expressing the respective TSHR constructs before and after partial purification. FIG. 1c shows Western blotting analysis of the C-del TSHR expressed in High Five cells during different stages of partial purification;

FIGS. 2a and b are graphs illustrating the results of HPLC gel filtration; FIG. 2c is a photograph of the results of SDS-PAGE electrophoresis following the production of the TSHR 260-M22 Fab complex;

FIG. 3 is a stereo-view representation of a 2F_o-F_c electron density map showing the residues of TSHR-M22 Fab complex binding interface;

FIG. 4 is a diagram illustrating the secondary structure of the TSHR LRD shown in JOY format (Mizuguchi K, Deane C M, Blundell T L, Johnson M S, Overington J P 1998 JOY: protein sequence-structure representation and analysis. Bioinformatics 14: 617-623)(the TSHR LRD amino acid sequence is SEQ ID NO:14);

FIGS. 5a-e is a series of diagrams illustrating the interaction of M22 Fab with the TSHR:

FIG. 5a illustrates the molecular surface of the TSHR-M22 Fab complex;

FIG. 5b is an opened up view of the interface area with residues that are within 4 Å of each other highlighted and labeled;

FIG. 5c shows hypervariable regions of M22 Fab which are highlighted in different shades and labeled for clarity;

FIG. 5d shows the electrostatic potential surface of TSHR and M22 Fab;

FIG. 5e is a list of residues of M22 Fab hypervariable regions. L1 corresponds to the M22 Fab hypervariable region of the amino acid sequence of SEQ ID NO: 22. L2 corresponds to the M22 Fab hypervariable region of the amino acid sequence of SEQ ID NO: 23. L3 corresponds to the M22 Fab hypervariable region of the amino acid sequence of SEQ ID NO: 24. H1 corresponds to the M22 Fab hypervariable region of the amino acid sequence of SEQ ID NO: 25. H2 corresponds to the M22 Fab hypervariable region of the amino acid sequence of SEQ ID NO: 26. H3 corresponds to the M22 Fab hypervariable region of the amino acid sequence of SEQ ID NO: 27;

FIG. 6a is a diagrammatic superposition of FSHR-FSH complex structure with TSHR-M22 Fab complex structure;

FIG. 6b shows the representation from FIG. 6a rotated clockwise 90 degrees about the vertical axis;

FIG. 6c is a structure based sequence alignment of TSHR and FSHR in JOY format (Mizuguchi K, Deane C M, Blundell T L, Johnson M S, Overington J P 1998 JOY: protein sequence-structure representation and analysis. Bioinformatics 14: 617-623) (the FSHR amino acid sequence is SEQ ID NO:15);

FIG. 6d is a spacefill representation of contact surfaces of TSHR LRD with M22 Fab and of FSHR LRD with FSH. The amino acids involved in the interactions are shown in darker grey colour;

FIG. 7 is a schematic diagram of the amino acid residues interacting across the interface of the TSHR-M22 Fab complex;

FIG. 8 is a stereo view of direct interactions (distances<4 Å) observed at the antibody-receptor interface in the TSHR-M22 Fab complex;

FIG. 9a-b are tables of co-ordinates derived from crystallography experiments described below, with:

FIG. 9a=2.55 Å resolution listing;

FIG. 9b=3.1 Å resolution listing (Chain A=human thyroid stimulating autoantibody M22 light chain, chain B=human thyroid stimulating autoantibody M22 heavy chain, chain C=human thyrotropin receptor (TSHR), fragment=leucine rich repeat domain (segment 22-260). In the M22 co-ordinates the light chain leucine in position 1 and the heavy chain threonine in position 131 are shown as alanine. In the TSHR co-ordinates glutamic acid in position 34, Glutamic acid in position 35, aspartic acid in position 36; phenylalanine in position 37, lysine in position 42, arginine in position 112 are shown as alanine. These residues were modeled as alanine or glycine due to lack of electron density);

FIG. 10a illustrates the consensus amino acid sequence (SEQ ID NO:16) of TSHR (accession no. P16473);

FIG. 10b shows the sequence (SEQ ID NO:17) identified by M Misrahi, H Loosfelt, M Atger, S Sar, A Guiochon-Mantel, E Milgrom. Cloning, sequencing and expression of human TSH receptor. Biochem Biophys Res Commun 1990 166: 394-403 (accession no. M32215);

FIG. 10c shows the sequence (SEQ ID NO:18) identified by A L Frazier, LS Robbins, P J Stork, R Sprengel, D L Segaloff, R D Cone. Isolation of TSH and LH/CG receptor cDNAs from human thyroid: regulation by tissue specific splicing. Mol Endocrinol 1990 4: 1264-1276 (accession no. M73747);

FIG. 10d shows the sequence (SEQ ID NO:19) identified by Y Nagayama, K D Kaufman, P Seto, B Rapoport. Molecular cloning, sequence and functional expression of the cDNA for the human thyrotropin receptor Biochem Biophys Res Commun 1989 165: 1184-1190 (accession no. M31774);

FIG. 10e shows a reference sequence (SEQ ID NO:20) (accession no. NM_—000369); and

FIG. 10f shows the sequence (SEQ ID NO:21) given in EP 0 433 509A.

DISCLOSURE

Production of TSHR260 in Insect Cells

The TSHR260 construct (coding for amino acids 1-260 of the human TSHR shown in FIG. 10a) was amplified using the full length human TSHR as the template (Oda Y, Sanders J, Roberts S, Maruyama M, Kato R, Perez M, Petersen V B, Wedlock N, Furmaniak J, Rees Smith B 1998 Binding characteristics of antibodies to the TSH receptor. Journal of Molecular Endocrinology 20: 233-244) with 5′-cactgcaggatccaaatgaggccggcggacttg-3′ (SEQ ID NO:1) and 5′-cagtcctctagattatcagt gatggtggtggtgatggttaagagtccaggtgtttcttgctat-3′ (SEQ ID NO:2) primers (Sigma Genosys) which added a BamHI restriction site to the N terminus, a 1 amino acid linker (Asparagine) and a 6 histidine tag, stop codon and an XbaI restriction site to the C terminus. The PCR reaction was carried out for 25 cycles of 1 min 95° C., 1 minute 50° C. and 1 minute 72° C. and the TSHR 260 cloned into pFastbac1 (Invitrogen, UK) using BamHI and XbaI restriction sites and the DNA sequence verified by the Sanger Coulson method (Sanger F, Nicklen S, Coulson A R 1977 DNA sequencing with chain terminating inhibitors. Proceedings of the National Academy of Sciences of the USA 74: 5463-5467). Recombinant Bacmid DNA was made using the Bac to Bac Baculovirus expression system (Invitrogen, UK). Briefly, 1 ng (5 μL) of pFastBac1-TSHR 260 was transfected into 100 μL MAX efficiency DH10Bac cells (Invitrogen, UK) containing the bacmid (baculovirus shuttle vector) and a helper plasmid. After incubation on ice for 30 minutes the cells were heat shocked at 42° C. for 45 seconds then chilled on ice for 2 minutes before addition of 900 μL of SOC medium (Invitrogen, UK). The tubes were incubated with shaking at 37° C. for 4 hours then 10 fold serial dilutions in SOC medium were plated onto LB agar plates (Tryptone 10 g/L, yeast extract 5 g/L, NaCl 10 g/L and 20 g/L agar) containing 50 μg/mL kanamycin, 7 μg/mL gentamycin, 10 μg/ml tetracycline, 100 μg/mL X-gal (Promega, UK) and 40 μg/mL isopropyl-3-D-thiogalactopyranoside (IPTG) followed by incubation at 37° C. for 48 hours to allow blue/white colour selection. White colonies were grown, recombinant plasmid DNA prepared using a plasmid midi kit (Qiagen, UK) and the presence of the TSHR260 DNA in the recombinant Bacmid confirmed using PCR.

The recombinant Bacmid DNA was transfected into Sf-9 insect cells (Invitrogen, UK), grown in TC100 medium (Invitrogen, UK) supplemented with 10%(v/v) fetal calf serum and 7 μg/ml gentamycin, to obtain and amplify recombinant baculovirus stock. Virus stocks were harvested from SF-9 cultures by centrifugation at 500 g for 5 minutes and retaining the supernatant that was stored at 2-8° C. All virus stocks were titered using a BacPAK baculovirus rapid titer kit (BD Clonetech) according to the manufacturer's instructions.

In addition, two further TSHR constructs for expression in the insect cell system were prepared as described above. The TSHR277 construct (coding for amino acids 1-277 of the human TSHR shown in FIG. 10a) was amplified using the full length human TSHR as the template using primers: 5′-caggaaacagctatgac-3′ (SEQ ID NO:3) and 5′-gctactcgagctagtggtggtggtggtggtgaaggtcagcccgtgtgaggtgaaggaaactcaag-3′ (SEQ ID NO:4) (Sigma Genosys) which added a 6 histidine tag, stop codon and an XhoI restriction site to the C terminus. The PCR reaction was carried out for 25 cycles of 1 minute 95° C., 1 minute 40° C. and 1 minute 72° C. and the TSHR 277 cloned into pFastbac1 (Invitrogen, UK) using BamHI and XhoI restriction sites and the DNA sequence verified by the Sanger Coulson method (Sanger F, Nicklen S, Coulson A R 1977 DNA sequencing with chain terminating inhibitors. Proceedings of the National Academy of Sciences of the USA 74: 5463-5467). Recombinant Bacmid DNA was prepared as for TSHR260 and the presence of recombinant Bacmid confirmed by PCR. Recombinant Bacmid DNA was transfected into Sf-9 insect cells, virus stock prepared and titered as described above for the TSHR260 construct.

The C-del TSHR construct coding for the TSHR amino acids 1-410 except for the sequence coding for the TSHR amino acids 313-353 which was deleted from the sequence (the TSHR amino acids were as shown in FIG. 10a) was also produced. In addition, the TSHR C-del construct contained three mutations ie R312E, E358T and E360T introduced in order to prevent proteolytic cleavage. The construct was made in four separate stages. Firstly the double mutation E358T1E360T was introduced using the full length TSHR as template. Two separate PCR reactions were set up (PCR 1 and PCR 2) as described previously in WO2006/016121A. PCR1 reaction used the T7 promoter primer 5′-taatacgactcactataggg-3′ (SEQ ID NO:5) and “reverse” primer for mutation 5′-accaatgatctcatccgtttgtgatcaaagaagacgta-3′ (SEQ ID NO:6) while PCR2 reaction used the “forward” primer for mutation 5′-tacgtcttctttgaaacacaaacggatgagatcattggt-3′ (SEQ ID NO:7) and the bovine growth hormone polyadenylation signal reverse primer (BGHR) 5′-tagaaggcacagtcgagg-3′ (SEQ ID NO:8). The PCR1 and 2 reactions were carried out using a GeneAmp PCR System 9700 (Applied Biosystems) at 94° C. for 5 min followed by 30 cycles of 94° C. for 1 min, 40° C. for 1 min and 72° C. for 2 min. PCR1 and PCR 2 products were excised from agarose gels and cleaned using a Geneclean II kit (Anachem Ltd, Luton, LU2 0EB, UK) according to the manufacturer's instructions. Purified PCR1 and PCR2 products were used to set up PCR 3 to construct the whole TSHR sequence containing the mutation. The PCR 3 reactions contained 200 ng of PCR 1 product and 200 ng of PCR 2 product. PCR 3 was carried out for 7 cycles of 94° C. 1.5 minutes, 65° C. 1.5 minutes and 72° C. for 1.5 minutes. The temperature was then increased to 94° C. again for 2 minutes and the T7 primer and BGHR primers added followed by 30 cycles of 94° C. 1 minute, 52° C. 1 minute and 72° C. 2 minutes. The PCR 3 product (TSHR E358T/E360T) was cloned into the pcDNA 5.1/FRT vector (Invitrogen) using BamHI and XhoI restriction sites and the presence of the mutation was verified using sequencing by the Sanger-Coulson method (Sanger F, Nicklen 5, Coulson A R 1977 DNA sequencing with chain terminating inhibitors. Proceedings of the National Academy of Sciences of the USA 74: 5463-5467).

The TSHR E358T/E360T construct was then used as the template DNA to introduce the third mutation (R312E) using the same method as above for the TSHR E358T/E360T. The PCR 1 reaction was carried out using the “reverse” primer for mutation 5′-gcattcacagatttttcctggcgcaagctctgca-3′ (SEQ ID NO:9) while PCR2 reaction used the “forward” primer for mutation 5′-tgcagagcttgcgccaggaaaaatctgtgaatgc-3′ (SEQ ID NO:10). Purified PCR1 and PCR2 products were used to set up PCR 3 as described above and the PCR 3 product (TSHR E358T/E360T/R312E) was cloned into the pcDNA 5.1/FRT vector (Invitrogen) using BamHI and XhoI restriction sites and the presence of the mutation was verified using sequencing by the Sanger-Coulson method (Sanger F, Nicklen S, Coulson A R 1977 DNA sequencing with chain terminating inhibitors. Proceedings of the National Academy of Sciences of the USA 74: 5463-5467).

The TSHR E358T/E360T/R312E (containing triple mutation) was then used as template DNA to delete amino acids 313-353 by the PCR method described above. The PCR 1 reaction was carried out using the “reverse” primer for deletion 5′-catccgtugtgatcaaagaagacttcctggcgcaagctagcatactg-3′ (SEQ ID NO:11) while PCR2 reaction used the “forward” primer for deletion 5′-cagtatgcagagcttgcgccaggaagtcttetttgaaacacaaacggatg-3′ (SEQ ID NO:12). Purified PCR1 and PCR2 products were used to set up PCR 3 as described above and the C-del TSHR product coding for TSHR amino acids 1-764 with the residues 313-353 deleted was cloned into the pcDNA 5.1/FRT vector (Invitrogen) using BamHI and XhoI restriction sites and the deletion verified using sequencing by the Sanger-Coulson method (Sanger F, Nicklen 5, Coulson A R 1977 DNA sequencing with chain terminating inhibitors. Proceedings of the National Academy of Sciences of the USA 74: 5463-5467) and then used as a template for amplification with the T7 promoter primer and 5′-gctactcgagctagtggtggtggtggtggtggtcttcacacgggttgaactcatcggacttg-3′ (SEQ ID NO:13) which added a 6 histidine tag, a stop codon and XhoI restriction site to the C terminus after TSHR amino acid 410. The PCR reaction was carried out for 25 cycles of 1 minute 95° C., 1 minute 50° C. and 1 minute 72° C. and the C-del TSHR cloned into pFastbac1 (Invitrogen, UK) using BamHI and XhoI restriction sites and the DNA sequence verified by the Sanger Coulson method (Sanger F, Nicklen S, Coulson A R 1977 DNA sequencing with chain terminating inhibitors. Proceedings of the National Academy of Sciences of the USA 74: 5463-5467). Recombinant Bacmid DNA was prepared as for TSHR260 and the presence of recombinant Bacmid confirmed by PCR. Recombinant Bacmid DNA was transfected into Sf9 insect cells, virus stock prepared and titered as described above for the TSHR260 construct.

Preparation of Purified M22 IgG and Fab

M22 IgG was prepared from heterohybridoma culture supernatants using protein A affinity chromatography on MabSelect™ (GE Healthcare, UK) as described in: Sanders J, Jeffreys J, Depraetere H, Evans M, Richards T, Kiddie A, Brereton K, Premawardhana L D K E, Chirgadze D Y, Núñez Miguel R, Blundell T L, Furmaniak J, Rees Smith B 2004 Characteristics of a human monoclonal autoantibody to the thyrotropin receptor: sequence structure and function. Thyroid 14: 560-570. The purified IgG was treated with mercuripapain (Sigma, UK) at an enzyme/protein ratio of 1:50 and passed through a MabSelect™ column to remove any intact IgG or Fc from the Fab preparation as described in: Sanders J, Jeffreys J, Depraetere H, Evans M, Richards T, Kiddie A, Brereton K, Premawardhana L D K E, Chirgadze D Y, Núñez Miguel R, Blundell T L, Furmaniak J, Rees Smith B 2004 Characteristics of a human monoclonal autoantibody to the thyrotropin receptor: sequence structure and function. Thyroid 14: 560-570. M22 Fab was analyzed on SDS polyacrylamide gel electrophoresis (SDS-PAGE) under reducing conditions to assess purity. M22 Fab biological activity was tested in cyclic AMP stimulation assays using Chinese hamster ovary (CHO) cells expressing TSHRs. In addition, the ability of M22 Fab to inhibit ¹²⁵I-TSH or ¹²⁵I-M22 binding to TSHR coated tubes was also analyzed (Sanders J, Oda Y, Roberts S, Kiddie A, Richards T, Bolton J, McGrath V, Walters S, Jaskolski D, Furmaniak J, Rees Smith B 1999 The interaction of TSH receptor autoantibodies with ¹²⁵I-labeled TSH receptor. Journal of Clinical Endocrinology and Metabolism 84: 3797-3802).

Production of the TSHR 260-M22 Fab Complexes

High Five™ insect cells (BTI-TN-5B1-4 from Invitrogen, UK) were maintained in ExCell 400 medium (SAFC Biosciences) supplemented with 0.1 mmol/L KI and 7 μg/mL gentamycin, at 22° C. in 500 mL spinner flasks (Techne), stirring at 60 rpm with ventilation. Each flask was seeded at a cell density of 0.5×10⁶ cells/mL and incubated for 24 hours at 22° C. before infecting with baculovirus stock at multiplicity of infection (MOI) of 0.0006 pfu/cell. Incubation of cell cultures was continued and purified M22 Fab was added 96 hours post-infection to a final concentration of 2 μg/mL. Culture supernatant containing the TSHR260-M22 Fab complex was harvested 120 hours post-infection by centrifugation at 500 g for 10 minutes. One tablet of Complete protease inhibitors (Roche) was added per 200 mL of supernatant, before storing at −70° C. until purification.

A separate series of experiments was carried out to assess the stability of TSHR260 during production in High Five insect cell cultures in the absence of M22 IgG, in the presence of 8 μg/mL M22 IgG and in the presence of 12 μg/mL M22 IgG on days 3, 4, 5 and 6 post infection. M22 IgG was added to the High Five cell culture supernatants on the day of infection. The integrity of the expressed TSHR260 was determined by Western blotting analysis of the respective samples of the culture supernatants (FIG. 1a). Western blotting was carried out using a mouse monoclonal antibody reactive with a TSHR epitope within amino acids 246-260 (TSHR MAb 18C5) (Jeffreys J, Depraetere H, Sanders J, Oda Y, Evans M, Kiddie A, Richards T, Furmaniak J, Rees Smith B 2002 Characterization of the thyrotropin binding pocket. Thyroid 12: 1051-1061) at 1 μg/mL concentration.

Specifically, FIG. 1a shows the TSHR260 detected in High Five cell culture supernatants in the absence or in the presence of M22 IgG on day 3, 4, 5 and 6 post infection (lanes 1-4, respectively). Panel 1=Samples of cell culture supernatants from High Five cells expressing the TSHR260 in the absence of M22 IgG (lanes 1, 2, 3 and 4 represent days 3, 4, 5 and 6 post infection respectively and lane 5 is High Five culture supernatants from cells not expressing the TSHR260 used as a negative control). Panel 2=Samples of cell culture supernatants from High Five cells expressing the TSHR260 in the presence of 8 μg/mL M22 IgG (lanes 1, 2, 3 and 4 represent days 3, 4, 5 and 6 post infection respectively and lane 5 is High Five culture supernatants from cells not expressing the TSHR260 used as a negative control). Panel 3=Samples of cell culture supernatants from High Five cells expressing the TSHR260 in the presence of 12 μg/mL M22 IgG (lanes 1, 2, 3 and 4 represent days 3, 4, 5 and 6 post infection respectively and lane 5 is High Five culture supernatants from cells not expressing the TSHR260 used as a negative control).

As shown in FIG. 1a, the intensity of the band of molecular weight 34 kDa representing the TSHR 260 increases to a maximum on day 5 post infection in cultures without M22 IgG. However, on day 6 post infection (panel 1; lane 4) most of the TSHR260 product has degraded to a protein of molecular weight 29 kDa or formed an aggregated band of 41 kDa. A similar pattern was observed when 8 μg/mL of M22 IgG was added into the culture media although on day 6 more of the intact TSHR260 was present in addition to degraded or aggregated material compared to the experiments without M22 IgG (FIG. 1a panel 2). When the concentration of M22 IgG in the culture media was increased to 12 μg/mL, a similar amount of the intact TSHR260 was present on day 6 to that detected on day 5 (FIG. 1a panel 3).

These results show that the TSHR260 is protected in the culture media by forming a complex with M22 IgG. The stability of the TSHR277 and the C-del TSHR products expressed by High Five insect cells in the presence of M22 Fab (added to the culture media on day 4 post infection to a final concentration of 2 μg/mL) was studied in further experiments. The culture supernatants from High Five cells infected with the virus carrying the respective TSHR construct were harvested on day 5 post infection and partially purified using chromatography on Streamline HST matrix (as described below). The presence and the molecular weight of the expressed TSHR products were determined by Western blotting analysis of the respective samples of the culture supernatants (FIGS. 1b and c). Western blotting was carried out using a mouse monoclonal antibody reactive with a TSHR epitope within amino acids 246-260 (TSHR MAb 18C5) (Jeffreys J, Depraetere H, Sanders J, Oda Y, Evans M, Kiddie A, Richards T, Furmaniak J, Rees Smith B 2002 Characterization of the thyrotropin binding pocket. Thyroid 12: 1051-1061) at 1 μg/mL concentration.

Specifically; FIG. 1b shows the TSHR277, C-del TSHR and TSHR260 present in the harvested culture supernatants before purification and after partial purification on the Streamline column. Lane 1=sample of cell culture supernatant from the High Five cells expressing the TSHR277 and lane 2=the partially purified TSHR277. Lane 3=the C-del TSHR present in cell culture supernatants and the same material after partial purification shown in lane 4. Lanes 5 and 6=the TSHR260 before and after partial purification, respectively. Some non-specific binding of TSHR MAb 18C5 to M22 Fab (at molecular weight 46 kDa) was observed in lanes 2 and 4.

As shown in the FIG. 1b, the TSHR277 of approx. molecular weight 34 kDa was expressed by High Five cells cultured in the presence of M22 Fab however, after partial purification the molecular weight of the TSHR277 decreased to 30 kDa (FIG. 1b lanes 1 and 2). The C-del TSHR expressed in High Five cells in the presence of M22 Fab was detected in a sample of cell culture supernatant as a 46 kDa protein band which degraded to a 30 kDa protein after partial purification (FIG. 1b lanes 3 and 4). The TSHR260 in this series of experiments was detectable as 30-32 kDa protein before and after purification (FIG. 1b lanes 5 and 6).

This experiment shows that although the products of the expected molecular weight for the TSHR277 and the C-del TSHR were expressed in High Five cultures (34 kDa and 46 kDa, respectively) both proteins degraded after partial purification to the size of the TSHR260 protein (approximately 30 kDa).

A further example of degradation of the C-del TSHR during partial purification is shown in FIG. 1c. The C-del TSHRs were expressed in the presence of 2 μg/mL of M22 Fab in High Five cell culture.

Specifically, FIG. 1e shows the C-del TSHR in High Five cell culture supernatants and during different stages of purification on a Streamline HST matrix. Lane 1=the TSHR260 in High Five cell culture supernatant; lane 2=the C-del TSHR in High Five cell culture supernatant; lane 3=cell culture supernatant from High Five cells expressing the C-del TSHR diluted and adjusted for load onto a Streamline HST column; lane 4=Streamline HST column flow through material; lane 5-8=eluted Streamline HST column fractions containing partially purified C-del TSHR.

As shown in FIG. 1c, the C-del TSHR of molecular weight 46 kDa was expressed into the culture supernatant by High Five cells grown in the presence of M22 Fab (lane 2). The intact C-del TSHR was also detectable in a load material for a Streamline HST column (lane 3) and no C-del TSHR was detectable in the column flow through (lane 4). However, the C-del TSHR eluted from a column has degraded to about 30 kDa molecular weight (lanes 5-8) which is comparable to the molecular weight of the TSHR260 shown in lane 1.

This series of experiments indicates that the TSHR polypeptide chain of different length as expected from the constructs used are expressed in High Five cells in the presence of M22 Fab. However, even during early stages of purification the TSHR 277 and the C-del TSHR products degraded to the size of the TSHR260 product. There was no evidence of degradation of the TSHR260 during purification (also see below). It is most likely therefore, that binding of M22 Fab involves large parts of the TSHR260 polypeptide chain and protects it from proteolysis. The TSHR amino acid sequence C-terminal from the residue 260 is unlikely to be involved in a stable binding of M22 Fab and consequently the TSHR sequences between amino acids 260 and 277 or 260 and 410 are subject to proteolytic degradation.

TSHR260-M22 Fab complex was found to be surprisingly stable. This is in contrast to previous TSHR polypeptide preparations which have been unstable, denaturing quickly under production conditions. The TSHR260-M22 Fab complex was analyzed, after two rounds of affinity purification before and after deglycosylation with Endoglycosidase F3, on HPLC gel filtration and by SDS-PAGE electrophoresis as shown in FIGS. 2a-c.

Specifically, FIG. 2a shows: purified TSHR260-M22 Fab complex before deglycosylation-analysis by gel filtration HPLC (TSK-GEK G3000SW run in 150 mmol/L NaCl, 10 mmol/L Tris pH 7.0; fraction volume 0.5 mL). Peak 1=TSHR260-M22 Fab complex. Peak 2=M22 Fab alone superimposed from a run carried out separately. FIG. 2b shows purified TSHR260-M22 Fab complex after deglycosylation, separation on cation exchange HPLC and concentration; ie material used for crystallisation-analysis by gel filtration HPLC as FIG. 2a. Peak 1=TSHR260-M22 Fab complex, peak 2=free M22 Fab. In FIG. 2c an analysis of purified TSHR260-M22 Fab by SDS-PAGE (12% acrylamide gel) under non-reducing conditions with the positions of molecular weight markers is shown. The positions of M22 Fab and TSHR260 before and after deglycosylation are marked. Lane 1=TSHR260-M22 Fab before deglycosylation; Lane 2=after deglycosylation and before purification on cation exchange HPLC column; Lane 3=after deglycosylation and purification on cation exchange HPLC column; Lane 4=after deglycosylation and purification on cation exchange HPLC column and final concentration. Approximately 10 μg of TSHR260-M22 Fab were loaded per lane in lanes 1, 2 and 4 and approximately 5 μg per lane in lane 3.

As shown in FIG. 2a, the TSHR260-M22 Fab complex ran as a sharp, almost homogeneous peak (peak 1, FIG. 2a) with only a small amount (5%) of free M22 Fab present in the preparation (FIG. 2a). After deglycosylation and separation on a cation exchange HPLC column the integrity of TSHR260-M22 complex as judged by gel filtration HPLC was intact (peak 1, FIG. 2b) and only a small amount of free M22 Fab (7%; peak 2, FIG. 2b) was detected in the final preparation used for crystallisation.

Analysis of the purified TSHR260-M22 Fab by SDS-PAGE under non-reducing conditions resolved the complex into its two components (TSHR260 and M22 Fab) in approximately equal proportions (lane 1, FIG. 2c). When calculated from SDS-PAGE, the molecular weight of glycosylated M22 Fab was approximately 56 kDa whereas the molecular weight of TSHR260 was approximately 40 kDa. The samples after deglycosylation (before separation from endoglycosidase F3 on cation exchange HPLC, after cation exchange HPLC, and after concentration; FIG. 2c, lanes 2, 3 and 4 respectively) all showed similar protein band patterns. The top band of approximately 56 kDa present in the smallest proportion represented some remaining non-deglycosylated M22 Fab. The protein band of molecular weight approximately 54 kDa represented deglycosylated M22 Fab while the band at molecular weight of approximately 37 kDa represented deglycosylated TSHR260.

Analysis of the N-terminal amino acids of the protein band at molecular weight approximately 40 kDa (as shown in FIG. 2c, lane 1) revealed the following sequence: Met-Gly-X-Ser-Ser-Pro. The X is likely to be Cys. This corresponds to the sequence of human TSHR between amino acids 22-27 ie Met-Gly-Cys-Ser-Ser-Pro and is consistent with the expressed sequence starting with residue 22 (residues 1-21 are the signal peptide) (Misrahi M, Loosfelt H, Atger M, Sar S, Guiochon-Mantel A, Milgrom E 1990 Cloning, sequencing and expression of human TSH receptor. Biochemical and Biophysical Research Communications 166: 394-403). Consequently, the identity of TSHR260 in our purified complex was confirmed by N-terminal amino acid sequencing.

Purification of TSHR260-M22 Fab Complex

Culture supernatant (in two batches of 14.4 and 17.4 L) was adjusted to pH 6.2 with 500 mmol/L NaH₂PO₄ and loaded onto 75 mL of Streamline Direct HST matrix in a Streamline 25 expanded bed chromatography system (GE Healthcare, UK). A further batch of culture supernatant (11.9 L) was processed in the same way in a separate experiment. The column was washed with 50 mmol/L sodium phosphate pH 6.0, 50 mmol/L NaCl, followed by 100 mmol/L NaCl, 50 mmol/L Tris-HCl pH 6.5 and elution with 100 mmol/L NaCl, 50 mmol/L Tris-HCl pH 8.0. The presence of the TSHR260-M22 complex in eluted fractions was confirmed by Western blotting analysis using a mouse monoclonal antibody reactive with a TSHR epitope within amino acids 246-260 (TSHR MAb 18C5) (Jeffreys J, Depraetere H, Sanders J, Oda Y, Evans M, Kiddie A, Richards T, Furmaniak J, Rees Smith B 2002 Characterization of the thyrotropin binding pocket. Thyroid 12: 1051-1061) at 1 μg/mL concentration.

The TSHR260-M22 Fab complex was further purified by affinity chromatography using an antibody TSHR MAb 14C4 (Jeffreys J, Depraetere H, Sanders J, Oda Y, Evans M, Kiddie A, Richards T, Furmaniak J, Rees Smith B 2002 Characterization of the thyrotropin binding pocket. Thyroid 12: 1051-1061) that binds to a conformational epitope within amino acids 22-261 of the TSHR extracellular domain, coupled to CNBr-activated sepharose-4B (Sigma, UK). Briefly, the complex was loaded onto a 4 mL affinity column, washed with 100 mmol/L NaCl, 50 mmol/L Tris-HCl pH 8.0, eluted with 100 mmol/L NaCl, 100 mmol/L citrate pH 4.0 and collected into an equal volume of neutralisation buffer (0.5 mol/L Tris-HCl pH 8.0) followed by dialysis into 50 mmol/L NaCl, 10 mmol/L Tris-HCl pH 8.0.

The dialysed complex was then further purified using Nickel affinity chromatography. The complex was loaded onto a Ni-NTA agarose column (Qiagen, UK), washed with wash buffer (50 mmol/L NaCl, 10 mmol/L Tris-HCl pH 8.0) and the complex eluted with 20 mmol/L immidazole in wash buffer. The complex was dialysed into 50 mmol/L NaCl, 10 mmol/L Tris-HCl pH 8.0 and used to set up deglycosylation reactions. The concentration of the complex was calculated from the absorbance at 280 nm on the basis that 1 absorbance unit is equivalent to 0.69 mg/mL of TSHR260-M22 Fab.

Deglycosylation of the TSHR 260-M22 Fab Complex and Final Purification

16 mg of purified complex obtained from 31.8 L of culture supernatant (or 7.5 mg obtained from 11.9 L culture supernatant in a further experiment) was deglycosylated using Endoglycosidase F3 (Sigma, UK) at an enzyme to complex ratio of 152 mU/mg complex in 50 mmol/L sodium acetate buffer pH 4.5 at 20 C for 5 days. The deglycosylation reactions were applied onto a cation exchange HPLC Bioassist S column (TOSOH). Briefly, the reactions were diluted 1:1 with 200 mmol/L Tris-HCl pH 6.8, filtered through a 0.22 μm filter before loading onto the column. The column was then washed with 20 mmol/L NaCl, 20 mmol/L NaCl pH 6.5 before elution of the complex using a pH gradient from pH 6.5 to pH 9.0. The purified complex (5 mg, or 2.4 mg in a further experiment) was then concentrated to 32.8 mg/mL (or 32 mg/mL) using a Microcon YM-10 concentrator (Millipore, UK), analyzed by gel filtration using an HPLC TSK gel G3000SW column (TOSOH) to determine the integrity of the complex and analyzed by SDS-PAGE to assess the purity. This material was used for crystallization screening trials or stored at −20° C. in aliquots.

Protein Sequencing of TSHR 260

Purified TSHR260-M22 Fab complex (30 μg) was run on a 12% SDS-PAGE (under non-reducing conditions) followed by blotting onto Immobilon P^SQ transfer membrane (Millipore, Watford, UK) in 10 mmol/L 3[cyclohexylamino]-1-propanesulfonic acid (CAPS) (pH 11) and 10% methanol. The membrane was stained with Coomassie blue, and the band representing TSHR260 excised and the N-terminal amino acid sequence analyzed (Alta Biosciences, Birmingham, UK).

Amino Acid Mutations in the M22 Heavy Chain (HC) or Light Chain (LC)

The M22 HC and LC sequences with ‘C’ terminal six histidine tags were cloned into vectors derived from pUC 18 as described in Sanders J, Jeffreys J, Depraetere H, Evans M, Richards T, Kiddie A, Brereton K, Premawardhana L D K E, Chirgadze D Y, Núñez Miguel R, Blundell T L, Furmaniak J, Rees Smith B 2004 Characteristics of a human monoclonal autoantibody to the thyrotropin receptor: sequence structure and function. Thyroid 14: 560-570.

Specific “forward” and “reverse” PCR primers were designed for each mutation to change the nucleotide coding sequence of either the HC or LC to code for the appropriate amino acid mutation. The primers were made by Sigma Genosys (Cambridge, UK). Two separate PCR reactions were set up (PCR 1 and PCR 2). In the case of the LC PCR1 reactions the M13 reverse sequencing primer and “reverse” primer for mutation were used while PCR2 reactions used the “forward” primer for mutation and the −20 M13 “forward” sequencing primer. For the HC PCR1 reactions the M13 “reverse” sequencing primer and “forward” primer for mutation were used while PCR2 reactions used the “reverse” primer for mutation and the −20 M13 “forward” sequencing primer. The PCR1 and PCR2 reactions were carried out using a GeneAmp PCR System 9700 (Applied Biosystems, UK) at 94° C. for 5 minutes followed by 30 cycles of 94° C. for 1 minute, 42° C. for 1 minute and 72° C. for 2 minutes. PCR1 and PCR2 products were excised from agarose gels and cleaned using a Geneclean II kit (Anachem Ltd, UK) according to the manufacturer's instructions. Purified PCR1 and PCR2 products were then used to set up PCR3 to construct the whole HC or LC sequences containing the mutation. The PCR3 reactions contained 200 ng of PCR 1 product and 200 ng of PCR 2 product. PCR3 was carried out for 7 cycles of 94° C. 1.5 minutes, 65° C. 1.5 minutes and 72° C. for 1.5 minutes. The temperature was then increased to 94° C. again for 2 minutes and the −20 M13 “forward” sequencing primer and the M13 “reverse” sequencing primer were added followed by 30 cycles of 94° C. 1 minute, 42° C. 1 minute and 72° C. 2 minutes.

The wild type or mutated M22 HC were cloned into the XhoI and SpeI restriction sites and the wild type or mutated M22 LC were cloned into the Sad and XbaI restriction sites of the Immunozap H/L vector (Stratagene Europe; Amsterdam, Netherlands) and the presence of the mutation verified using sequencing by the Sanger-Coulson method as described in the art.

Plasmid DNA containing the M22 HC and LC sequences was transformed into HB2151 cells (Amersham Biosciences) and pre-cultures; one colony of transformed HB2151 in 10 mL of LB ampicillin (Tryptone 10 g/L, Yeast Extract 5 g/L, NaCl 10 g/L, Ampicillin 100 μg/mL) containing 1% glucose were grown overnight at 30° C. with shaking. Thereafter the pre-cultures were diluted (5 mL in 500 mL LB ampicillin) and grown at 30° C. until the absorbance at 600 nm was 1.2. Then 1.8 mol/L sucrose was added to a final concentration of 0.3 mol/L and the cultures incubated at 30° C. until the absorbance at 600 nm returned to 1.2. Thereafter isopropyl-β-D thiogalactoside (IPTG) was added to a final concentration of 1 mmol/L and cultures continued for 24 hours at 23° C. with shaking. The cultures were then centrifuged at 9000 rpm for 30 minutes at 4° C., 1 mmol/L phenylmethylsulfonyl fluoride (PMSF) and 1 Complete protease inhibitor tablet (Roche) per 100 mL of supernatant added and the supernatant stored at −70° C. before purification. Expression of recombinant Fab was verified using Western blotting analysis with an antihuman IgG (Fab specific) antibody (Sigma, UK). In Western blotting analysis 10 ng of wild type recombinant M22 was detectable.

Purification of Recombinant Wild Type and Mutated M22 Fab

The supernatants (4 liters used for each purification) containing recombinant M22 Fab were adjusted to pH 6.0 with 500 mmol/L sodium dihydrogen phosphate pH 4.0 and loaded onto a 75 mL Streamline Direct HST column (GE Healthcare, UK). The column was washed with 10 mmol/L Tris-HCl pH 6.8, 0.1 mol/L NaCl until an absorbance at 280 nm was below 0.1 and M22 Fab eluted with 0.3 mol/L NaCl and 10 mmol/L Tris-HCl pH 8.3. The eluted material was loaded onto a Ni-NTA agarose column (Qiagen, UK), washed with 0.3 mol/L NaCl and 10 mmol/L Tris-HCl pH 8.3 containing 40 mmol/L imidazole followed by 120 mmol/L imidazole for elution of M22 Fab.

The purity of the eluted Fabs was >95% as assessed by SDS-PAGE and the concentration of the Fabs was calculated from the absorbance at 280 nm on the basis that 1 absorbance unit is equivalent to 0.7 mg/mL of Fab.

Amino Acid Mutations in the Human TSHR Sequence

The method used to introduce specific mutations into the TSHR sequence was as described in WO2006/1016121A.

Briefly, the TSHR full length nucleotide sequence was cloned into the pcDNA5.1/FRT vector (Invitrogen) using BamHI and XhoI restriction sites following standard cloning procedures. Specific “forward” and “reverse” PCR primers were designed for each mutation to change the nucleotide coding sequence to code for the appropriate amino acid mutation. All primers were made by Sigma Genosys (Cambridge, UK). Two separate PCR reactions were set up (PCR 1 and PCR 2). PCR1 reactions used the T7 and “reverse” primer for mutation while PCR2 reactions used the “forward” primer for mutation and the bovine growth hormone polyadenylation signal reverse primer (BGHR primer). The PCR 1 and 2 reactions were carried out using a GeneAmp PCR System 9700 (Applied Biosystems) at 94° C. for 5 min followed by 30 cycles of 94° C. for 1 min, 40° C. for 1 min and 72° C. for 2 min. PCR1 and PCR 2 products were excised from agarose gels and cleaned using a Geneclean H kit (Anachem Ltd, Luton, LU2 0EB, UK) according to the manufacturer's instructions. Purified PCR1 and PCR2 products were used to set up PCR 3 to construct the whole TSHR sequence containing the mutation. The PCR 3 reactions contained 200 ng of PCR 1 product and 200 ng of PCR 2 product. PCR 3 was carried out for 7 cycles of 94° C. 1.5 min, 65° C. 1.5 min and 72° C. for 1.5 min. The temperature was then increased to 94° C. again for 2 min and the T7 primer and BGHR primers added followed by 30 cycles of 94° C. 1 min, 52° C. 1 min and 72° C. 2 min. The PCR 3 product was cloned into the pcDNA 5.1/FRT vector (Invitrogen) using BamHI/XhoI restriction sites and the presence of the mutation was verified using sequencing by the Sanger-Coulson method as described in the art.

Transfection of Mutated TSHR Constructs into CHO Cells Using the Flp-In System

The method used was as described in WO2006/016121A. Briefly, the Flp-In system (Invitrogen) was used for transfection of wild type (Wt) and mutated TSHR cDNAs into CHO cells. The Flp-In-CHO cells contain one Flp-In site per cell and consequently TSHR DNAs will be inserted in the same place in the genome in each experiment and will be present only as one copy per cell. A confluent flask of Flp-In-CHO cells was used to seed 24 well plate wells at 1×10⁵-1.5×10⁵ cells/well in DMEM (Invitrogen), 10% fetal calf serum (FCS) (Invitrogen), with no antibiotics and the cells were incubated overnight at 37° C., 5% CO2 and >95% humidity. The pcDNA5.1/FRT TSHR DNA and POG44 DNA (Invitrogen) were then transfected into the Flp-In CHO cells using lipofectamine (Invitrogen) according to the manufacturer's instructions. The cells were selected using 600 μg/mL of hygromycin (Invitrogen) in the media with those cells transfected with both POG44 plasmid DNA and pcDNA5.1/FRT TSHR being capable of inserting the TSHR into the Flp-In-CHO cell genome and showing hygromycin resistance.

Analysis of Stimulation of Cyclic AMP Production

The method used to measure stimulation of cyclic AMP production in CHO cells transfected with wild type or mutated TSHRs, in order to confirm functionality of the constructs, is described in detail in WO2006/016121A.

Briefly, CHO cells were seeded into 96 well plates (12,500-20,000 cells per well) and incubated for 48 hours in DMEM containing 10% fetal calf serum. The DMEM was then removed and dilutions of porcine TSH (RSR Ltd; 0.01-3 ng/mL) and wild type or mutated M22 Fab (0.1-10 ng/mL) in cyclic AMP assay buffer (NaCl free Hank's Buffered Salts solution containing 1 g/L glucose, 20 mmol/L HEPES, 222 mmol/L sucrose, 15 g/L bovine serum albumin (BSA) and 0.5 mmol/L 3 isobutyl-1-methyl xanthine, pH 7.4) were added and incubated for 1 hour at 37° C. in an atmosphere of 5% CO₂ in air. After removal of the test solutions, cells were lysed and cyclic AMP concentrations in the lysates determined using a Biotrak enzyme immunoassay system from Amersham Biosciences.

Preparation of Detergent Solubilised Wild Type and Mutated TSHR Preparations

The procedure used was as described in WO2006/016121A.

Flp-In-CHO cells expressing either the wild type or mutated TSHR were grown to confluence in 175 cm² flasks, the cells washed with Dulbecco's PBS (without calcium and magnesium ions) (Invitrogen) and scraped into 10 mL ice cold buffer A (50 mmol/L NaCl, 10 mmol/L Tris-HCl pH 7.5), containing 1 tablet of Complete protease inhibitors (Roche Diagnostics) per 200 mL of buffer and 1 mmol/L phenylmethanesulfonyl fluoride). The cells were pelleted at 1000×g for 5 min at 4° C., the pellet resuspended in 1 mL buffer A and homogenised in a glass homogeniser on ice. The cell membranes were pelleted at 12,000×g for 30 min at 4° C. and resuspended in 6 mL of buffer A plus 0.5 g/L sodium azide and 2.75 g/L iodoacetamide and pelleted as above. The membrane pellet was then resuspended in 1 mL ice cold buffer A containing 1% Triton X-100 and 0.5 g/L sodium azide and homogenized. The solubilized TSHR preparations were centrifuged at 90,000×g for 2 hours at 4° C. and the supernatants stored at −70° C. in aliquots.

Inhibition of ¹²⁵I-M22 IgG or ¹²⁵I-TSH Binding to the TSHR

¹²⁵I-labeled M22 IgG or ¹²⁵I-labeled TSH binding inhibition assays were carried out using tubes coated with wild type TSHR as described previously (Sanders J, Oda Y, Roberts S, Kiddie A, Richards T, Bolton J, McGrath V, Walters S, Jaskolski D, Furmaniak J, Rees Smith B 1999 The interaction of TSH receptor autoantibodies with ¹²⁵I-labeled TSH receptor. Journal of Clinical Endocrinology and Metabolism 84: 3797-3802) (reagents from RSR Ltd). A calibration curve prepared using M22 Fab prepared from IgG purified from M22 hybridoma culture supernatants (1-100 ng/mL) was included in each assay.

In the assay, 100 μL of purified wild type or mutated Fab (0.001-100 μg/mL diluted in assay buffer (50 mmol/L NaCl, 10 mmol/L Tris-HCl pH 7.8, 0.1% Triton X-100)) were incubated in TSHR coated tubes at room temperature for 2 hours with gentle shaking. After aspiration, the tubes were washed twice with 1 mL of assay buffer before addition of 100 μL of ¹²⁵I-M22 IgG (50,000 cpm) or ¹²⁵I-TSH (80,000 cpm) and incubation at room temperature for 1 hour with shaking. The tubes were then washed twice with 1 mL of assay buffer, aspirated and counted in a gamma counter.

Inhibition of M22 IgG or TSH binding was calculated as:

$100\times 1-\frac{\mathrm{cpm}M22\mathrm{or}\mathrm{TSH}\mathrm{bound}\mathrm{in}\mathrm{the}\mathrm{presence}\mathrm{of}\mathrm{test}\mathrm{material}}{\left(\mathrm{cpm}\mathrm{bound}\mathrm{in}\mathrm{the}\mathrm{presence}\mathrm{of}\mathrm{assay}\mathrm{buffer}\right)}$
Crystallization and Diffraction Data Collection

De-glycosylated TSHR-M22 Fab complex at a concentration of 32.8 mg/mL was used for vapour-diffusion hanging-drop crystallisation experiments. Clusters of small crystals appeared after about two weeks in Wizard Crystal Screen 1, condition #46 (Emerald BioStructures, Inc.). The crystallisation solution was then optimised to 8% PEG8000, 0.1 mol/L MES pH 6.0, 0.25 mol/L zinc acetate, which resulted in bigger crystals but still growing in clusters. Attempts to use micro/macro-seeding techniques to obtain single crystals were unsuccessful. A single crystal with dimensions 0.02×0.02×0.05 mm³ was manually separated from a cluster and flash-cooled in liquid nitrogen in the presence of 26% ethylene glycol as the cryo-protectant agent.

X-ray diffraction data collection experiments were performed at 100K using an “in-house” copper-rotating anode radiation source (generator RU-H3R, Rigaku-MSC Ltd. equipped with Max-Flux confocal multilayer optics, Osmic Inc.). The diffraction data were recorded using Raxis IV++ image plate detector (Rigaku-MCS Ltd.). The raw diffraction data were collected using a single crystal at one degree oscillation steps (a total of 129 degrees were collected) and were indexed, integrated, scaled and reduced using HKL diffraction data processing suite (Otwinowski Z, Minor W 1997 Processing of X-ray diffraction data collected in oscillation mode. Methods in Enzymology: Macromolecular Crystallography, part A 276: 307-326). The crystal belonged to the orthorhombic I2₁2₁2₁ space group, had one TSHR-M22 Fab complex in the asymmetric unit (54% solvent content) and diffracted to 3.1 Å Bragg's spacing. The refinement statistics are shown in Table 1.

A further complex of the TSHR260 with M22 Fab was prepared as described, concentrated to 32 mg/mL and set up for the crystallisation trials and crystals were obtained with the same crystallization conditions as the first series of experiments. X-ray diffraction data collection was then carried out using a synchrotron radiation source (station PX14.2, Council for the Central Laboratory of the Research Councils, Daresbury, UK). The data were collected from a single crystal at 100 K and the resolution of diffraction was improved to 2.55 Å Bragg's spacing. The initial structure obtained at 3.1 Å resolution was then refined again using the newly acquired 2.55 Å resolution data to an R-factor of 18.1% (R_free=24.5%). The refinement statistics are shown in Table 1.

RESULTS

Structure Determination and Refinement

Partially deglycosylated TSHR260-M22 Fab complex was successfully crystallized but only produced clusters of multiple crystals which could not be developed further in order to obtain a single crystal. A single crystal suitable for X-ray diffraction analysis was obtained by manually splitting the cluster of crystals.

The structure was solved by molecular replacement. M22 Fab (Sanders J, Jeffreys J, Depraetere H, Evans M, Richards T, Kiddie A, Brereton K, Premawardhana L D K E, Chirgadze D Y, Núñez Miguel R, Blundell T L, Furmaniak J, Rees Smith 13 2004 Characteristics of a human monoclonal autoantibody to the thyrotropin receptor: sequence structure and function. Thyroid 14: 560-570) and FSHR (Fan Q R, Hendrickson W A 2005 Structure of human follicle-stimulating hormone in complex with its receptor. Nature 433: 269-277) crystal structures were used as the molecular replacement search models; the calculations were done in AMoRe (Navaza J 1994 Amore—an automated package for molecular replacement. Acta Crystallography Section D 50: 157-163) of CCP4 program suite (Bailey S 1994 The CCP4 suite—programs for protein crystallography. Acta Crystallography Section D 50: 760-763). The M22 Fab search probe was further split into two: one, containing only variable domains and the other—constant domains. The positions of TSHR and variable domains of M22 Fab within the asymmetric unit were successfully obtained, resulting in an R-factor of 48.2%. However, no solution could be identified for the constant domains, these were subsequently placed manually using electron density maps calculated after a preliminary refinement round. The resulting model, prior to the refinement, had an R-factor of 43.5% and R_free of 45.5%. A total of eight rounds of crystallographic refinement and manual rebuilding were performed. The atomic crystallographic refinement was done using CNS (Brunger A T, Adams P D, Clore G M, DeLano W L, Gros P, Grosse-Kunstleve R W, Jiang J S, Kuszewski J, Nigles M, Pannu N S, Read R J, Rice L M, Simonson T, Warren G L 1998 Crystallography and NMR system: a new software suite for macromolecular structure determination. Acta Crystallography Section D 54: 905-921) and, at the later stages, REFMAC (Murshudov G N, Vagin A A, Dodson E J 1997 Refinement of macromolecular structures by the maximum-likelihood method. Acta Crystallography Section D 53: 240-255) packages. Simulated annealing protocols as implemented in CNS were used in the first two rounds of refinement, but were replaced by the Powell minimization protocol in the last rounds. Manual rebuilding was performed in Coot (Emsley P, Cowtan K 2004 Coot: model-building tools for assessing the accuracy of crystal structures. Nature 355: 472-475) using sigmaA weighted 2F_o-F_c, F_o-F_c and annealed omit maps. The Zn²⁺ ions, N-acetylglucosamine residues and water molecules were only placed in the last refinement/rebuilding rounds. The model of the complex structure was refined at 3.1 Å resolution to an R-factor of 20.7% (R_free=28.3%). The initial structure obtained at 3.1 A resolution was then refined again using the newly acquired 2.55 Å resolution data to an R-factor of 18.1% (R_free=24.5%) (Table 1).

The final structure consists of M22 LC (residues 1-208), HC (residues 1-127 and 134-213), TSHR (residues 30-257), six N-acetylglucosamine residues, 5 zinc ions and 289 water molecules. Continuous electron density was observed at all N-linked glycosylation sites on TSHR (N77, N99, N113, N177 and N198) and the one site on M22 (LC N26). There was no electron density for the loop residues of M22 HC 128-133 and some terminal residues (M22 Fab LC 209-212, M22 Fab HC 214-220, TSHR residues 22-29, 258-260 and the C terminal hexa-histidines) due to disorder. Side-chain atoms of TSHR E35 were lacking clear electron density and therefore this residue was modeled as alanine.

For an example of the electron density map, see FIG. 3. Specifically in FIG. 3, part of the interface between the TSHR (generally towards the top of the figure) and M22 Fab heavy chain (generally towards the bottom of the figure) is shown. The map is contoured at 1.2 σ level, all residues displayed are labeled.

TSHR Structure

The TSHR has the shape of a slightly curved tube, having opposed concave and convex surfaces, with a ten-stranded β-sheet located on the concave surface. The inner surface of the tube is lined with hydrophobic residues. The closest homologue of TSHR is FSHR with which it shares 40.9% sequence identity (Misrahi M, Loosfelt H, Atger M, Sar S, Guiochon-Mantel A, Milgrom E 1990 Cloning, sequencing and expression of human TSH receptor. Biochemical and Biophysical Research Communications 166: 394-403 (FIG. 10b) and Fan Q R, Hendrickson W A 2005 Structure of human follicle-stimulating hormone in complex with its receptor. Nature 433: 269-277); the root mean square deviation (rmsd) on C_α core atoms between the structures is 1.1 Å. The concave surface of the leucine rich repeat structure of the human TSHR presents ten β strands in one parallel β sheet and forming nine repeats. The number of residues in each strand from the N-terminus is: 4, 5, 5, 5, 5, 7, 5, 6, 3, 3. The additional β strand before the strand of the first repeat forms a β hairpin. There are eight small strands (two residues each) in the convex surface of the structure of the LRD forming two, three-stranded 3 sheets and one, two-stranded β sheet. There are no helices in the TSHR LRD structure as can be seen in FIG. 4. Secondary structures were identified using SSTRUC software developed by David Keith Smith (1989, unpublished data) based on the DSSP algorithm (Kabsch W, Sander C 1983 Dictionary of protein secondary structure: pattern recognition of hydrogen-bonded and geometrical features. Biopolymers 22:2677-2637). All five (N77, N99, N113, N177 and N198) glycosylation sites on the TSHR are located on the convex surface.

TSHR-M22Fab Complex

The structure of the TSHR-M22 Fab complex shows M22 Fab bound to the concave surface of TSHR260 with an axis of symmetry along the TSHR “tube” nearly parallel to the interface between the light and heavy chains of the autoantibody as shown in FIG. 5. The majority of the residues located on the binding interface of the antibody variable regions of M22 when bound to the TSHR have almost identical positions compared to those in un-bound M22 (rmsd of all atoms 0.4 Å). The highest deviation of an atom from M22 backbone residues is only 1.1 Å observed for C_α atom of HC P97. In addition, only six M22 residues present a deviation of their side chains compared to un-bound M22 greater than 2 Å (Table 5).

The overall position of M22 Fab constant domains is different from that seen in the unbound M22 Fab structure by about 20 degrees of rotation around the axis between the constant and variable domains due to the packing of molecules in the crystal. There was adequate electron density to model carbohydrate residues in all five potential glycosylation sites on the receptor and one (N26) on M22 Fab. All glycosylation sites on TSHR260 are far away from the binding interface and do not interfere with M22 Fab binding. Comparison of TSHR260-M22 Fab complex with the complex of FSH LRD with FSH is shown in FIG. 6. In FIG. 6, a cartoon diagram of both structures is shown, the superposition was performed using FSHR and TSHR residues only.

Some aspects of the binding arrangements in the complex are particularly surprising. These include:

• • (a) M22 clasps the concave surface of the TSHR LRD in a very similar manner to the way FSH clasps the FSH LRD. Furthermore the 2 fold axis of M22 and the 2 fold axis of FSH in their respective complexes overlap completely. It is remarkable that an autoantibody adopts almost identical binding features to the hormone. There is no hint of this in any of the prior art.
  • (b) In addition the area of the concave surface of the TSHR LRD which interacts with M22 is large (2500 Ų) and extends from the N to the C terminus. This is surprising and unexpected and there are no hints in the prior art of such extensive interactions.
  • (c) In addition to the large area of binding with the LRD, the strength and number of different types of interactions observed in the crystal structure of the complex is surprising. For example, to have a network of 22 hydrogen bonds and salt bridges between two interacting proteins is most unusual.
  • (d) Comparison of the structure of M22 in the complex and free M22 shows that essentially no movement in the M22 atoms of residues involved in TSHR binding occurred. Again this is surprising as some induced fit would have been expected. Also there is considerable movement in FSH when it binds to the FSHR. With regard to the structure of the TSHR LRD itself, this was surprisingly similar to the FSHR LRD but with some important differences including the number of β strands.
    Autoantibody-Receptor Interactions

A total of 2,500 Ų of solvent-accessible surface area is buried in the interface between the autoantibody and the receptor. The interactions between TSHR and M22 Fab represent a mixture of an extensive hydrogen bonding network, salt bridges (22 hydrogen bonds and salt bridges), non-hydrogen bonding polar interactions and hydrophobic contacts (Table 2; FIGS. 7 and 8). In FIG. 8 interacting residues are shown as sticks and labeled. Hydrogen bonds are shown as dotted lines.

The heavy chain of M22 Fab has more residues than the light chain which interact with the TSHR, although both chains form a number of hydrogen bonds and salt bridges (14 for the heavy chain and 8 for the light chain). The majority of interacting residues of M22 Fab are located in hypervariable regions L2, H1, H2 and H3 as defined by Kabat (Kabat E, Ferry H, Wu T, Gottesman K, Foeller C 1991 Sequences of proteins of immunological interest, 5th ed. US Public Service Health Service. Bethesda, Md.) as can be seen in FIG. 5e. In FIG. 5e, residues involved in receptor binding are in bold and underlined. (Otwinowski Z, Minor W 1997 Processing of X-ray diffraction data collected in oscillation mode. Methods in Enzymology: Macromolecular Crystallography, part A 276: 307-326). The surface buried in the interface between TSHR LRD and M22 LC is 526.9 Ų for TSHR LRD and 526.5 Ų for M22 LC whereas for the interaction between the TSHR LRD and M22 HC the areas are 730.1 Ų for TSHR LRD and 730.4 Ų for M22 HC.

There are 7 hydrogen bonds between the TSHR LRD and M22 LC and 7 hydrogen bonds between the TSHR LRD and M22 HC. In particular, M22 HC Y99 is hydrogen bonded with two TSHR residues; E107 (involving the backbone nitrogen of M22 Y99 and the side chain of TSHR E107) and with K58 (involving the side chains of both residues). Also M22 LC Q53 produces two hydrogen bonds (with TSHR N208 and Q235). TSHR K129 produces three hydrogen bonds involving M22 HC T30 and T53 while TSHR Q235 is hydrogen bonded to two M22 residues (LC D52 and LC Q53). The TSHR260-M22 Fab interactions also include 14 water mediated hydrogen bonds (Table 2). The TSHR residues involved in strong van der Waals interactions with M22 (ie an interaction surface area of more than 60 Ų) are R255 (102.2 Ų), R80 (91.9 Ų), R 38 (76.6 Ų), K129 (75.5 Ų), K183 (64.4 Ų) and R109 (60.6 Ų). The M22 residues involved in strong van der Waals interactions with TSHR260 (interaction surface area of more than 70 Ų) are: HC R28 (115.8 Ų), HC Y56 (86.6 Ų), LC Y50 (86.4 Ų), HC Y99 (79.4 Ų), LC Q53 (76.7 Ų) and HC P97 (70.8 Ų). The electrostatic interactions present in the TSHR260-M22 Fab complex involve the following residues (in order of decreasing strength): TSHR D151 and M22 HC R28 (minimum distance 2.76 Å), TSHR K58 and M22 LC D95A (minimum distance 2.60 Å), TSHR R80 and M22 HC D54 (minimum distance 2.75 Å), TSHR K183 and M22 HC E96 (minimum distance 3.04 Å), TSHR K209 and M22 LC D52 (minimum distance 3.57 Å), TSHR R80 and M22 HC D52 (minimum distance 3.20 Å), TSHR K129 and M22 HC R28 (minimum distance 4.05 Å), TSHR K209 and M22 LC D51 (minimum distance 4.50 Å), TSHR R255 and M22 LC D60 (minimum distance 4.39 Å), TSHR K129 and M22 HC D52 (minimum distance 5.27 Å)(Table 3) determined by the Henderson-Hasselbalch equation which was used to calculate the charges of the side chains of residues taking the pH into consideration as implemented in an in house program for assessing atomic charges and the distances between charged atoms.

Out of the TSHR residues, TSHR R80 produces the strongest accumulated electrostatic interactions with M22 residues while M22 HC R28 produces the strongest accumulated electrostatic interactions with TSHR residues. The residues of the hypervariable regions H1, H2 and H3 of M22 Fab heavy chain form an outer edge of the negatively charged cavity which interacts with a highly positively charged area of TSHR (R38, K58, R80, H105, K129).

Effect of Mutations in the TSHR LRD on M22 Stimulation of Cyclic AMP Production in TSHR Expressing CHO Cells in Relation to Interactions Seen in the Crystal Structure of TSHR260-M22 Fab

The experimentally determined effects of amino acid mutations in the TSHR LRD on M22 stimulating cyclic AMP activity is described in WO2006/016121A. These effects were then analyzed in view of the interactions found in the crystal structure of TSHR260-M22.

For example, mutations R80A, E107A, R109A, K129A, K183A, Y185A, R255A had a marked effect on M22 stimulating activity (below 60% of activity with the wild type TSHR) (WO2006/016121A) and analysis of the interactions between TSHR260 and M22 Fab in the crystal structure of the complex indicated that all these TSHR residues interact with M22 Fab (Tables 2 and 3). Furthermore, they are involved in 9 out of the 22 hydrogen bonds and salt bridges present in the structure (Table 2) while TSHR R109 produces two water mediated hydrogen bonds (Table 2) and strong van der Waals interactions. Also, R80, E107, R109, K183, Y185 and R255 are involved in both non-hydrogen bonding polar interactions and in hydrophobic contacts with M22 (Table 2).

One other mutation in the TSHR LRD that had a marked effect on M22 stimulating activity as described in WO2006/016121A was F130A and the crystal structure shows F130 in hydrophobic contacts with M22 HC P97 and HC G98 (Table 2).

Several new mutations in the TSHR LRD (not described in WO2006/016121A) were carried out and tested. Mutation K209E had a marked effect on M22 stimulating activity (<20% of activity with the wild type TSHR). TSHR K209 is involved in non-hydrogen bonding polar interactions with M22 LC Y50, forms hydrophobic contacts with M22 LC Q53 (Table 2) and attractive electrostatic interactions with M22 LC D51 and LC D52 (Table 4) and this provides an explanation why mutation TSHR K209E resulted in loss of M22 activity (<20% of activity with wild type TSHR).

Effect of Mutations in M22 on M22 Stimulation of Cyclic AMP in TSHR Transfected CHO Cells in Relation to Interactions Seen in the Crystal Structure of TSHR260-M22 Fab

Several single amino acid mutations were introduced into the M22 Fab heavy and light chain sequences and the effects of these mutations on M22 stimulation of cyclic AMP production in CHO cells expressing the wild type TSHR studied (Table 4).

Out of the several amino acids that showed effects on M22 stimulating activity (below 80% of wild type activity) when mutated, HC R28 forms three salt bridges with TSHR D151, polar interaction with TSHR F153 and 1152 and is in hydrophobic contact with TSHR 1152 and F153 (Table 2). M22 HC D52 is involved in strong electrostatic interactions with TSHR R80 and polar interactions with TSHR H1.05 (Tables 2 and 3). M22 HC D54 is in strong electrostatic interaction with TSHR R80 and is hydrogen bonded with TSHR T104 through water (Tables 2 and 3), M22 HC Y56 is involved in polar interactions with TSHR R38 and strong hydrophobic contacts with TSHR R38, T56 and R80 (Table 2). Finally, M22 LC D52 forms hydrogen bonds with TSHR Q235 and is in strong electrostatic interaction with TSHR K209 (Tables 2 and 3).

Combined analysis of the results of mutations in the TSHR and M22 on M22 stimulating activity together with analysis of the interactions between TSHR and M22 Fab in the crystal structure of their complex allowed identification of the residues in the TSHR and in M22 Fab that are important in interactions which result in biological activity. In particular, TSHR R38 is hydrogen bonded with M22 HC T57 while TSHR R80 forms 3 salt bridges with M22 Fab (HC D52 and HC D54) and is involved in hydrophobic contacts with M22 HC Y56 (Table 2). Further, TSHR K129 forms 3 hydrogen bonds with M22 Fab (HC T30 and HC T53). Consequently, the cluster of positively charged residues at the N-terminal end of the concave surface of the TSHR LRD is important in the interactions with M22 which result in biological activity (FIG. 5d).

However, residues in the C-terminal part of the concave surface of the TSHR LRD were also found to be important for M22 biological activity in our experiments and to be involved in interactions between the TSHR and M22 in the crystal structure of the complex. Of these residues, TSHR R255 is of particular interest as mutation of R255 has marked effect on M22 activity but has no effect on TSH activity (WO2006/016121A). TSHR R255 is involved in several interactions with M22 Fab in the crystal structure (two hydrogen bonds with M22 LC V58, electrostatic interactions with M22 LC D60, polar interactions with LC D60 and hydrophobic interactions with LC L54) (see above and Tables 2 and 3).

Many residues in both the heavy and the light chains of M22 are involved in interactions with the TSHR LRD in the complex (Table 2). Some of these, M22 HC R28, HC D52, HC D54, HC Y56 and LC D52 are of particular interest because as shown in our experiments, mutation of these residues had an effect on M22 biological activity (see above and Tables 2 and 4).

CONCLUSIONS

Overall, there was good agreement between analysis of the effects of various mutations in the TSHR or in M22 on M22-stimulating activity and the interactions observed in the crystal structure of TSHR260-M22 Fab. This indicates that the crystal structure of TSHR260 in complex with M22 Fab provides a means of designing molecular structures which will interact with the TSHR or interact with molecules like M22 in such a way as to interfere with the receptor-autoantibody interaction. Such interference provides a means of preventing the stimulatory effects of TSHR autoantibodies in patients with Graves' disease.

For example, a small molecule designed to fill up the negatively charged cavity on the M22 surface (formed by M22 H1, H2 and H3) which interacts with the highly positively charged ridge at the N-terminal end of the concave surface of the TSHR LRD should prevent M22 (and TSHR autoantibodies that have similar surface characteristics) binding to the receptor. Conversely, a small molecule designed to interact with the positively charged ridge on the TSHR LRD mentioned above would be expected to prevent M22 (and autoantibodies with similar surface properties) interacting with the TSHR. In addition, small molecules could be designed which prevent the critical TSHR residue R255 from interacting with M22 and other TSHR autoantibodies.

Specific amino acid mutations in the TSHR LRD and in M22 can be designed to study the mechanism of activation of the TSHR thereby indicating further means of preventing activation of the TSHR by TSHR autoantibodies. Furthermore, insights into the TSHR activation mechanism gained from these studies could provide means to investigate and understand gonadotropin receptor activation as gonadotropin receptors are closely related structurally to the TSHR.

Also, the detailed understanding of the interaction between M22 and the TSHR provided by the crystal structure of the complex in combination with further studies on the mechanism of receptor activation mentioned above will allow the design of new molecules which act as TSHR agonists. Such thyroid stimulating molecules would have application in vivo when tissue containing the TSHR needs to be stimulated. For example as an alternative to recombinant human TSH currently used to stimulate ¹³¹I uptake by any residual thyroid cancer left after ablative treatment.

The detailed information provided by the TSHR-M22 crystal structure will also allow the design of new and improved ligands for measuring and assessing TSHR autoantibodies in patient serum samples.

The interactions between the TSHR and M22 are extensive and complex. Furthermore, comparison of the crystal structure of the FSHR in complex with FSH (Fan Q R, Hendrickson W A 2005 Structure of human follicle-stimulating hormone in complex with its receptor. Nature 433: 269-277) and the TSHR-M22 crystal structure indicate that M22 positions itself relative to the TSHR in an almost identical way to the positioning of FSH relative to the FSHR (Table 6). In particular, both M22 and FSH clasp their respective receptors at about 90° to the receptor tube length axis. Comparative modeling of the TSH-TSHR interaction indicates that the complex formed by TSH and its receptor has a very similar structure to that formed by FSH in complex with the FSHR (Núñez Miguel R, Sanders J, Jeffreys J, Depraetere H, Evans M, Richards T, Blundell T L, Rees Smith B, Furmaniak J 2004 Analysis of the thyrotropin receptor-thyrotropin interaction by comparative modeling. Thyroid 14: 991-1011 and Núñez Miguel R, Sanders J, Blundell T L, Rees Smith B, Furmaniak J 2005 Comparative Modeling of the Thyrotropin Receptor. Thyroid 15: 746-747).

The evolutionary pressures on the immune system and on the TSHR which have resulted in the formation of autoantibodies which mimic the actions of TSH by interacting with the receptor in such a similar way to the hormone are intriguing. Now details of the M22-TSHR interaction are established definitively at the molecular level, some understanding of these evolutionary pressures and why TSHR autoimmunity has occurred may well become evident.

TABLE 1
Crystallographic data collection and refinement statistics at
2.55 Å resolution
X-ray diffraction data
Space group                      I212121
Unit cell: a, b, c (Å)           43.89, 175.78, 205.81
Resolution range (Å)             30.0-2.55 (2.61-2.55)
Rsym1 (%)                        7.1 (36.1)
Completeness (%)                 96.1 (99.2)
Number of unique reflections     25,731
Average redundancy               4.6
Average intensity, <I/σ(I)>      10.5
% reflections with I/σ(I) > 3 in the highest 47.5
resolution shell
Wilson B-factor (Å2)             47.7
Refinement
Resolution range (Å)             26.7-2.55
Number of reflections: work/test 23,125/1301
Rcryst2 (%)                      18.1
Rfree3 (%)                       24.5
Number of non-hydrogen atoms:
protein                          5,039
N-acetylglucosamine              84
Zn2+                             5
Water                            289
Model quality
Estimated coordinate error4 (Å)  0.30
Rms. deviation bonds (Å)         0.009
Rms. deviation angles (°)        1.236
Overall mean B-factor (Å2)       36.0
Ramachandran plot analysis5
Number of residues in:
allowed regions                  561
generously allowed regions       2
disallowed regions               2
Crystallographic data collection and refinement statistics at
3.1 Å resolution
X-ray diffraction data
Space group                      I212121
Unit cell: a, b, c (Å)           43.73, 175.16, 204.66
Resolution range (Å)             30.0-3.10 (3.17-3.10)
Rsym1(%)                         8.0 (40.4)
Completeness (%)                 99.4 (99.8)
Number of unique reflections     15,037
Average redundancy               4.2
Average intensity, <I/σ(I)>      8.7
% reflections with I/σ(I) > 3 in the highest 42.9
resolution shell
Wilson B-factor (Å2)             66.2
Refinement
Resolution range (Å)             28.3-3.10
Number of reflections: work/test 13,267/734
Rcryst2 (%)                      20.7
Rfree3 (%)                       28.3
Number of non-hydrogen atoms:
protein                          5,027
N-acetylglucosamine              84
Zn2+                             9
water                            38
Model quality
Estimated coordinate error4 (Å)  0.54
R.m.s. deviation bonds (Å)       0.006
R.m.s. deviation angles (°)      1.019
Overall mean B-factor (Å2)       46.0
Ramachandran plot analysis5
Number of residues in:
allowed regions                  558
generously allowed regions       4
disallowed regions               1
Values in parentheses show the corresponding statistics in the highest resolution shell.
1Rsym = Σh|Ih − <I>|/ΣhIh, where Ih is the intensity of reflection h, and <I> is the mean intensity of all symmetry-related reflections.
2Rcryst = Σ||Fobs| − |Fcalc||/Σ|Fobs|, Fobs and Fcalc are observed and calculated structure factor amplitudes.
3Rfree as for Rcryst using a random subset of the data (about 5%) excluded from the refinement (Brunger A T 1992 Free R value: a novel statistical quantity for assessing the accuracy of crystal structures. Nature 355: 472-475).
4Estimated coordinate error based on the Rfree value as calculated by REFMAC (Murshudov GN, Vagin AA, Dodson EJ 1997 Refinement of macromolecular structures by the maximum-likelihood method. Acta Crystallograpy Section D 53: 240-255).
5Calculated with PROCHECK (Laskowski R A, MacArthur M W, Moss D S, Thornton J M 1993 PROCHECK: a program to check the stereochemical quality of protein structures. J Appl Crystallogr 26: 283-291).

TABLE 2
Interactions between TSHR260 and M22 Fab observed in the
crystal structure of the complex
Hydrogen bonds and salt bridges1
TSHR	M22 Fab2	Distance, Å1
Arg38	NH1	Thr57 (B)	O	2.93
	NH2	Thr57 (B)	O	3.24
Lys58	NZ	Asp95A (A)	OD2	2.60 *
	NZ	Tyr99 (B)	OH	2.47
Arg80	NH1	Asp54 (B)	OD1	2.75 *
	NH1	Asp54 (B)	OD2	3.01 *
	NH2	Asp52 (B)	OD2	3.20 *
Glu107	OE2	Tyr99 (B)	N	2.64
Lys129	NZ	Thr30 (B)	OG1	2.80
	NZ	Thr30 (B)	O	2.78
	NZ	Thr53 (B)	OG1	2.82
Asp151	OD1	Arg28 (B)	NH1	3.38 *
	OD1	Arg28 (B)	NH2	2.76 *
	OD2	Arg28 (B)	NH1	2.87 *
Glu157	OE2	Tyr50 (A)	OH	2.50
Lys183	NZ	Glu96 (B)	OE1	3.04 *
Tyr185	OH	Tyr49 (A)	OH	2.74
Asn208	ND2	Gln53 (A)	OE1	3.06
Gln235	NE2	Asp52 (A)	OD2	3.16
	OE2	Gln53 (A)	NE2	2.95
Arg255	NH1	Val58 (A)	O	3.03
	NH2	Val58 (A)	O	2.81
Water-mediated hydrogen bonds
TSHR (distance to water, Å)	M22Fab (distance to water, Å)	Water
Arg38	NH2 (2.49)	Thr57 (B)	N (3.07)	201
Arg80	NH2 (2.34)	Asp52 (B)	OD2 (2.52)	27
Thr104	OG1 (2.63)	Asp54 (B)	OD2 (3.03)	66
His105	NE2 (2.80)	Thr30 (B)	O (2.85)	214
Glu107	OE1 (2.87)	Ser100 (B)	N (3.05)	100
Arg109	NH2 (3.04)	Tyr50 (A)	OH (3.18)	179
	NH2 (3.15)	Ser100 (B)	OG (3.01)	100
Lys129	O (2.95)	Ser31 (B)	OG (3.06)	44
Phe153	O (2.74)	Ser31 (B)	OG (3.06)	44
Glu157	OE1 (2.67)	Pro97 (B)	O (2.79)	68
Lys183	NZ (2.50)	Pro97 (B)	O (2.79)	68
Glu251	OE1 (3.00)	Ser56 (A)	OG (3.19)	209
Arg255	NH1 (2.96)	Leu54 (A)	O (2.80)	28
	NH2 (2.78)	Asp60 (A)	N (2.92)	43
TSHR	M22 Fab2
Non-hydrogen bonding polar interactions3
Arg38	Tyr56 (B)
His105	Asp52 (B)
Glu107	Gly98 (B)
Arg109	Ser100 (B)
Asn110	Asn30 (A)
Ile152	Arg28 (B)
Phe153	Arg28 (B), Ser31 (B)
Ile155	Tyr32 (B)
Lys183	Trp100C (B)
Tyr185	Tyr49 (A), Trp100C (B)
Asn208	Tyr49 (A)
Lys209	Tyr50 (A)
Glu251	Ser56 (A)
Arg255	Asp60 (A)
Asn256	Leu54 (A)
Hydrophobic contacts4
Arg38	Tyr56 (B)
Thr56	Tyr56 (B)
Lys58	Tyr99 (B)
Arg80	Tyr56 (B)
Tyr82	Tyr99 (B)
Phe130	Pro97 (B), Gly98 (B)
Ile152	Arg28 (B)
Phe153	Arg28 (B)
Ile155	Pro97 (B)
Tyr185	Tyr49 (A), Tyr50 (A)
Lys209	Gln53 (A)
Arg255	Leu54 (A)
* Denotes salt bridges.
1Hydrogen bond distances are in the range of 2.3-3.4 Å
2Letters in parenthesis indicate to which M22 Fab chain residues belong: A—light chain, B—heavy chain.
3Polar contacts have distances between 3.4 and 4.0 Å
4Carbon-carbon contacts are within 4.0 Å

TABLE 3
Ion pair interactions in the TSHR260-M22 Fab complex
	TSHR	M22 Fab
By residues (interaction of strength greater than 6.0e−10N)
	Lys58	LC Asp95A	(2.60 Å, 23.9e−10N)
	Arg80	HC Asp52	(3.20 Å, 15.3e−10N)
		HC Asp54	(2.75 Å, 18.4e−10N)
	Lys129	HC Arg28	(4.05 Å, 13.4e−10N)
		HC Asp52	(5.27 Å, 7.3e−10N)
	Asp151	HC Arg28	(2.76 Å, 24.1e−10N)
	Lys183	HC Glu96	(3.04 Å, 17.1e−10N)
	Lys209	LC Asp51	(4.50 Å, 8.8e−10N)
		LC Asp52	(3.57 Å, 16.8e−10N)
	Arg255	LC Asp60	(4.39 Å, 8.2e−10N)
By atoms (interaction of strength greater than 2.5e−10N)
	OD2 Asp36	NZ HC Lys64	(6.76 Å, 2.5e−10N)
	NH1 Arg38	NZ HC Lys64	(6.59 Å, 2.6e−10N)
	NH2 Arg38	NZ HC Lys64	(6.78 Å, 2.5e−10N)
	NZ Lys58	OD1 LC Asp95A	(4.08 Å, 6.9e−10N)
		OD2 LC Asp95A	(2.60 Å, 17.0e−10N)
	NH1 Arg80	OD1 HC Asp52	(4.62 Å, 2.7e−10N)
		OD2 HC Asp52	(3.44 Å, 4.9e−10N)
		OD1 HC Asp54	(2.75 Å, 7.6e−10N)
		OD2 HC Asp54	(3.01 Å, 6.4e−10N)
	NH2 Arg80	OD2 HC Asp52	(3.20 Å, 5.6e−10N)
	NZ Lys129	NH1 HC Arg28	(4.23 Å, 6.4e−10N)
		NH2 HC Arg28	(4.05 Å, 7.0e−10N)
		OD1 HC Asp52	(5.23 Å, 4.2e−10N)
		OD2 HC Asp52	(6.07 Å, 3.1e−10N)
		NZ HC Lys73	(7.17 Å, 4.5e−10N)
	OD1 Asp151	NH1 HC Arg28	(3.38 Å, 5.1e−10N)
		NH2 HC Arg28	(2.76 Å, 7.6e−10N)
	OD2 Asp151	NH1 HC Arg28	(2.87 Å, 7.0e−10N)
		NH2 HC Arg28	(3.58 Å, 4.5e−10N)
	NZ Lys183	OE1 HC Glu96	(3.04 Å, 12.4e−10N)
		OE2 HC Glu96	(4.97 Å, 4.7e−10N)
	NZ Lys 209	OD1 LC Asp51	(6.10 Å, 3.1e−10N)
		OD2 LC Asp51	(4.50 Å, 5.7e−10N)
		OD1 LC Asp52	(3.57 Å, 9.0e−10N)
		OD2 LC Asp52	(3.86 Å, 7.7e−10N)
		NH2 LC Arg66	(6.41 Å, 2.8e−10N)
	NH2 Arg255	OD1 LC Asp60	(4.70 Å, 2.6e−10N)
		OD2 LC Asp60	(4.39 Å, 3.0e−10N)
	The interaction strengths, shown for comparison, are in Newtons and are calculated using an in house program (ELECINT, R. Núñez Miguel, unpublished) taking ε = 1 for electrostatic field calculation and pH = 7.4 for the calculation of charges of side chain atoms of charged residues using the Henderson-Hasselbalch equation.
	Distances are between charged atoms.

TABLE 4
Effects of mutations in M22 on M22 stimulation of cyclic AMP
production in CHO cells expressing wild type TSHR
                            Stimulation of cyclic AMP
Mutated M22 Fab preparation production by M22 Fab
wild type                   +++++
HC R28D                     +++
HC T30A                     +++++
HC D52A                     ++
HC D52K                     −
HC D54R                     +
HC Y56A                     ++
HC K64E                     +++++
HC K73D                     +++++
HC R94E                     +++
HC E96A                     ++++
HC E96R                     no expression detected
LC D51K                     +++++
LC D52A                     +++
LC D52R                     no expression detected
LC D93R                     +++++
+++++ = wild type activity (100%),
++++ = <100-80% of wild type activity,
+++ = <80-60% of wild type activity,
++ = <60-40% of wild type activity,
+ = <40-20% of wild type activity,
− = <20% of wild type activity.
The effects of each mutation on the ability of M22 Fab to inhibit labelled M22 and inhibit labelled TSH binding to the TSHR paralleled the effects of stimulation on cyclic AMP.

TABLE 5
Deviations in atom positions in the structure of unbound-M22
and bound M22
Backbone
The only change that may be taken into consideration in the backbone is:
Pro97 HC, displacement of 1.1 Å of its Cα atom.
Side chains
Displacements (more than 2 Å):
Arg28 HC, displacement of 4.8 Å of its NH2 atom.
Trp33 HC, displacement of 4.0 Å of its NE1 atom.
Arg66 LC, displacement of 3.2 Å of its NH1 atom.
Lys64 HC, displacement of 3.2 Å of its NZ atom.
Asp95 LC, displacement of 2.2 Å of its OD1 atom.
Asp52 HC, displacement of 2.1 Å of its OD2 atom.
Displacements (more than 1.3 Å and less than 2 Å ):
Ser56 LC, displacement of 1.8 Å of its OG atom.
Tyr99 HC, displacement of 1.6 Å of its OH atom.
Asp60 LC, displacement of 1.4 Å of its OD2 atom.
Pro97 HC, displacement of 1.3 Å of its CB atom.

TABLE 6
Comparison of FSHR-FSH and TSHR-M22 complexes
A	FSHR or TSHR residues involved in van der Waals interactions with
	FSH or M22 Fab (HC = heavy chain; LC = light chain)
	respectively.
B	FSHR or TSHR residues involved in hydrogen bond interactions with
	FSH or M22 Fab (MC = heavy chain; LC = light chain)
	respectively.
C	FSHR or TSHR residues involved in strong ion pair interactions with
	FSH or M22 Fab (HC = heavy chain; LC = light chain)
	respectively.
Residue numbering across tables A-C corresponds to equivalent residues
in the FSHR and in the TSHR sequences.
A Van der Waals interactions
	FSHR	FSH	M22	TSHR
	residue	chain	chain	residue
			HC	35
			HC	38*
	33	β
	34	β
	50	β	HC	56
	52*	β	LC, HC	58
	54	β	LC	60
	55*	α, β
	56	α
	57	α
			HC	79
	74	α	HC	80*
	76	β	HC	82
	78	β
	79*	α, β	LC	85
	81*	α
			HC	104
	99	α	HC	105
	101*	α, β	HC	107
	103	β	LC, HC	109
	104	α, β	LC	110
	106*	α
	123	α	HC	129*
	124	α	HC	130
	126	α
			LC, HC	134
	129*	α, β
	130	α
	131	α
	145	α	HC	151
			HC	152
	146	β	HC	153*
	148	α	HC	155*
	150	α	LC, HC	157
	152	α, β	LC	159
	153*	α, β	LC	160
	155	α
	156	α
	172	β
	174	α, β
	176*	α	LC, HC	183*
	178	α	LC, HC	185*
	179*	α, β
	196	β
	197	β
			LC, HC	206
			LC	208
	202	β	LC	209*
	222*	β
			LC	232
			LC	234
			LC	235
	242	β
	243	β	LC	251
	245	β	LC	253
			LC	255*
			LC	256
*Residues that produce strong interactions (ΔASA > 40 Å2).
B Hydrogen bonds
FSHR	FSH	M22	TSHR
residue	chain	chain	residue
		HC2	38
34	β
		HC	58
79	α
99	α
		HC	107
		HC3	129
129	α
		LC	157
179	α, β
		LC	185
		LC	208
		LC2	235
		LC2	255
X3 the residue produces three hydrogen bonds with the corresponding
chain
C Ion pair interactions (interaction of strength greater than 4.0e−10N)
FSHR	FSH	M22	TSHR
residue	chain	chain	residue
		LC, HC	34
		HC	36
		LC, HC2	38
34	α
		LC	42
50	β
52	β	LC, HC	58
		LC	61
57	α
73	α
74	α, β	HC3	80
76	β
81	α2, β
99	β
101	α, β2	HC	107
103	β	LC	109
		HC3	129
		HC	151
146	α2, β2
150	α	HC	157
153	α2, β2	LC2	160
171	α, β
		HC2	183
179	α2, β2
196	β2
202	β	LC3	209
227	β
245	β
		LC	255
X2 denotes two ion pair interactions,
X3 denotes three interactions.

Claims

1. A method of identifying a chemical entity which will interact with Thyroid Stimulating Hormone Receptor (TSHR) using a computer system comprising:

(a) generating on a computer system a representation of a three-dimensional structure of a TSHR polypeptide, or a homologue of such a TSHR polypeptide using data stored in a database, where the data represents a three-dimensional structure of a TSHR polypeptide structure, or a homologue of such a TSHR polypeptide, corresponding to TSHR polypeptide amino acid coordinates of FIG. 9a or FIG. 9b with a root mean square deviation from the backbone atoms between 0 Å and 3.1 Å, wherein the coordinates of FIG. 9a or FIG. 9b are derived using X-ray crystallographic analysis of the structure of the complex formed by a TSHR polypeptide and a TSHR polypeptide bonding entity, wherein the X-ray crystallographic analysis identifies an interaction;

(b) receiving, at a processor associated with the database, a selection for at least one chemical entity as a candidate to interact with at least one amino acid of the representation of a three-dimensional structure of a TSHR polypeptide, or a homologue of such a TSHR polypeptide;

(c) identifying whether the selected chemical entity will interact with at least one amino acid of the three-dimensional structure of the TSHR polypeptide or homologue thereof by identifying a representation of at least one interaction between the three-dimensional structure of the TSHR polypeptide and the selected chemical entity, wherein said interaction comprises at least one of the following or a combination thereof: an interaction with at least one of the following TSHR amino acids: R38, K58, R80, H105, or K129; an interaction forming a hydrogen bond with at least one of the TSHR amino acids: K129, E107, K58, or Y185; an interaction forming van der Waals interactions with at least one of the TSHR amino acid residues: R255, R80, K129, R38, or K183; an interaction forming electrostatic interactions with at least one of the TSHR amino acid residues: D151, K58, K129, R80, K209, or K183; or an interaction forming ion pair interactions with TSHR amino acid residue K209.

2. The method according to claim 1, wherein the chemical entity is a TSHR agonist.

3. The method according to claim 1, wherein the chemical entity is a TSHR antagonist.

4. The method according to claim 1, wherein a chemical entity is identified which will interact by forming a hydrogen bond with at least one of the TSHR amino acids: K129, E107, K58 and Y185.

5. The method according to claim 1, wherein a chemical entity is identified which will interact by forming van der Waals interactions with at least one of the TSHR amino acid residues: 8255, R80, K129, R38 and K183.

6. The method according to claim 1, wherein a chemical entity is identified which will interact by forming electrostatic interactions with at least one of the TSHR amino acid residues: D151, K58, K129, R80, K209 and K183.

7. The method according to claim 1, wherein a chemical entity is identified which will interact by forming ion pair interactions with TSHR amino acid residue K209.

8. The method according to claim 1, wherein said chemical entity which may potentially interfere with the binding of autoantibodies to the TSHR, and the method further comprises identifying a chemical entity which interacts with the highly positively charged ridge at the N-terminal end of the concave surface of the TSHR leucine-rich domain.

9. The method according to claim 8, wherein the autoantibodies are thyroid stimulating autoantibodies.

10. The method according to claim 9, wherein the autoantibodies are TSH antagonists.

11. The method of claim 1, wherein said computer system includes data storage means including the data corresponding to the TSHR polypeptide amino acid coordinates of FIG. 9a or FIG. 9b.

12. The method of claim 1, wherein said method further comprises outputting, to a display coupled to the processor, a three-dimensional display of an interaction between the at least one chemical entity and the at least one amino acid of the TSHR polypeptide.

13. The method of claim 1, further comprising:

storing, in a database, the data representing a three-dimensional structure of a TSHR polypeptide structure, or a homologue of such a TSHR polypeptide, corresponding to TSHR polypeptide amino acid coordinates of FIG. 9a or FIG. 9b and data representing at least one three-dimensional structure of a chemical entity.

14. A method of identifying a chemical entity which will interact with Thyroid Stimulating Hormone Receptor (TSHR) using a computer system comprising:

(a) generating on a computer system a representation of a three-dimensional structure of a three-dimensional structure of a TSHR polypeptide structure, or a homologue of such a TSHR polypeptide, corresponding to TSHR polypeptide amino acid coordinates of FIG. 9a or FIG. 9b with a root mean square deviation from the backbone atoms between 0 Å and 4 Å and data representing at least one three-dimensional structure of a chemical entity;

(b) receiving, at a processor associated with the database, a selection for at least one chemical entity as a candidate that binds to or blocks binding to at least one amino acid of the representation of a three-dimensional structure of a TSHR polypeptide, or a homologue of such a TSHR polypeptide;

(c) outputting, to a display coupled to the processor, a three-dimensional display of an interaction between the at least one chemical entity and the at least one amino acid of the TSHR polypeptide; and

(d) identifying, via the display, whether the selected chemical entity that will interact with at least one amino acid of the three-dimensional structure of the TSHR polypeptide or homologue thereof by identifying a representation of at least one interaction between the three-dimensional structure of the TSHR polypeptide and the selected chemical entity, wherein said at least one amino acid of the TSHR polypeptide or homologue is R38, K58, R80, E107, K129, D151, K183, Y185, K209, or 8255.